Compare news coverage from diverse sources around the world. Take advantage of their Holiday Sale sale to get 40% off unlimited access by going to: ground.news/lilyalexandre. Sale ends December 31.
Lily, did you ever hear about Nonself-Emptiness? Or Anatta or Sunyatta? Search Nonself and AI or Emptiness (Biddhism) and AI, you might be surprised about the Buddhist view, I'd love to speak to yoi but I know your free time is valuable. Kindest regards, signed a Gay Buddhist layman lol
People becoming transvestigators for AI will definitely hurt neurodivergent people. If you're constantly on the lookout for unusual communication (especially a flat affect), you'll end up dismissing disabled people.
a few weeks ago there was a tweet going around from a person who got called out for using AI generated email responses, except they did not use any AI, they were just autistic... I wonder how many people who have naturally flat voices/expressions or who learned how to communicate by copying others are going to be unintentionally hurt by people trying to detect AI. sucks for everyone from every angle.
On top of that, the normalization of AI avatars essentially amplifies any "abnormal" behaviors, quirks, tics, or whatever else. I mean look at the thing about closing your mouth fully between sentences and sustaining eye contact. Is it wrong somehow, to not do those things? Does it demonstrate some kind of flaw or bad morals? Are we comfortable letting people who don't always make eye contact be driven even further from the perceived normal? When you average out everyone's behaviors you just... Erase all the one-off behaviors and moderately rare behaviors in favor of a thousand clones of the average. Don't even get me started on what this means for minority cultures
I'm autistic and a few months ago someone replied to one of my youtube comments saying that it was clearly AI generated. I think the problem was that each of my sentences was a logically complete unit? I don't know. I was using grammar too well or something.
Regardless of how much autistic writing and LLM-generated text actually resemble each other, I imagine such discrimination probably stems from the common public perceptions of what it means to be "robotic" shaped by mass media. Portrayals of sentient robots often go for a human-like personality that still exhibits "machine-like" traits like being heavily logical, flat affect, etc., thus it's easiest to model them after the kind of humans most expressing those traits. Since such humans tend to be on the autism spectrum, we get "autistic-coded" portrayals of AI in media, whether intentional or not, which shapes public perception of what an AI looks like. p.s. Yes, I am myself on the spectrum.
I don't know if the people on the receiving end of the comments were autistic but I've spotted tons of accusations on Reddit of comments being written by Chat GPT. It reminds me entirely too much of how right wing ideologues will accuse random women of being trans for not adhering to their conceptions of "real womanhood".
The possibility of AI-generated figures being more likable, more charismatic, more attractive than people are normally capable of -- like the software mentioned in the calliagnosia story -- reminds me of "supernormal stimuli." It's a phenomenon where a creature has evolved to respond to a particular stimuli in its natural habitat, but when something new is introduced to its habitat that exaggerates that stimuli, the response is exaggerated as well. Examples include especially attractive artificial eggs that birds will prefer over their own laid eggs (common in cases of brood parasitism), jewel beetles that will try in vain to mate with shiny glass bottles that resemble an attractive mate more than actual other beetles do, and snack engineers making entire professions out of creating food that our brains respond to as chemically perfect. If AI can not only actually crack the intangibles, that top 10%, of humanity but exponentially improve at capturing it, then we'll be looking at an explosion of supernormally stimulating humans in photo and video -- possibly already seen in not only the reading of AI-generated, perfectly average faces as more trustworthy than not, but also in the predictable applications of AI to sexual material. Like, you've SEEN the almost laughably disproportionate waifus these guys are generating.
Your comment reminded me of how quite frequently in animal behavior research, the researchers will create a fake version of the animal and place it nearby the biological animal to see what kinds of mating or territorial or nurturing or whatever behavior can be elicited. Humans clearly really aren't any different, and quite a lot of (especially incel techbro) humans are not just as easily able to be fooled by cheap facsimile, but almost tripping over themselves desiring to be so fooled.
wonder if this could lead to a consious inversion where the inherent distrust of supernomality as likley fake could lead to "abnoramlity" becoming attractive?
@famitory I think this is part of what makes punk culture compelling. It's basically disillusioned people making "ugly" art in protest of the pristine. Cruelty Squad is a good modern example.
I find it validating, since I'm becoming increasingly disillusioned with "humanity". AI is hilarious to me because it's so effortlessly revealing that what humans value about themselves aren't actually unique human qualities, which is something I've known to be true my entire life.
"I'd argue that corporations quietly gathering our personal data is always dehumanizing" -10:52 it hurts, that this is a statement that *needs* to be said. why do we trust an authority, that just want to *use* us as profit? This is how creativity, what makes us human, dies
@johncasey9544 mobility sales and service with A Telecom company T. XD Basically you call in because something is wrong with your bill. I explain why and try to sell you a new line. Saw all manner of nonsense.
The AI version of you was so upsetting (in a good, effective way) because it made me feel that like, something was wrong with you, the real human Lily? Like someone was just off camera pointing a gun at you or you'd been hit really hard in the head or something - it made me worry for you!
@@PhilosophyTube Oh I see. You have a close working relationship with fellow Nebula creators. And thus you treat them like family and support them even if they are smaller creators. With less reach than you. That is logical, and it makes sense. Its like how Greater sage-grouse who are related to each other, even if they are only distantly related, will help each other in their Lekking Courtship displays. As it is a "net good" to help those who we have blood or a relationship with. Even if it costs us as an individual in the short term. What is true for birds, is true for humans, I suppose.
Is it just me or is the AI avatar version of Lily deeply terrifying in an uncanny valley horror type way? The random body movements and out of place hand gestures combined with the otherwise fairly convincing voice make me feel like I'm watching an ultra-realistic simulation breaking down
Absolutely, lmao. It manages to capture every tiny detail needed to create one of the most terrifying depictions of a "human" I've ever seen. It's so inhuman, and it fails to fundamentally give off the impression that it isn't just an imitation, but the imitation is otherwise so close. The deep-fake was also weird
@@lily_lxndr I mean, I kind of enjoy the horror aspect to it actually hahah. terrible for intended purpose but possibly great for horror movies (like most AI!)
i'm usually immune to uncanny valley but that was hard to watch. i wanted to break through the screen and shake that simulation awake from its deep nightmare of nothingness.
Referring to the portion on audio: I'm a musician who works with mixing live and recorded audio regularly. I remember that clip from your video, and had actually found the audio quality strangely comforting. In a video essay about losing ourselves in insubstantial content, hearing you speak in a way that signaled your presence in a physical room felt not only grounding, but like an intentional choice. it's funny, because the reverb could be declared as a textbook mistake in the audio world, but I felt like it truly enhanced the point you were making. Glad you spoke about it and I'm glad you kept it in.
So interesting, I immediately heard the difference when Lily pointed out the loss of personality in the synthesised version. I also think that people pay too much attention to the audio quality and now we crave the sensation of being there with a person in a room, experiencing what they are conveying. That's why people love early band recordings/demos, even though they sound like absolute balls a lot of the time... People also try to mimic the sound of a 1920s room mic capturing a jazz band in their recordings - it's the feeling, nostalgia, presence that's evoked by the recordings as well as the music itself.
One of the things that worries me most about AI is how its erosion of trust affects our reactions to those outside the norm. I remember reading a story about an autistic professor who was accused of using Chat GPT to write their emails because they sounded flat and emotionless. It seems like people already perceived as unusual or socially different will be first and most affected by AI's appropriation of human social interaction. Like you mentioned with the AI-generated images affecting your perception of art, I worry that AI will only bolster the already strong boundaries around which traits and qualities of communication are normal, acceptable, and likeable.
Yes. Being autistic myself I don't really worry about IA stuff. Any language used to "convince" or “forced emotions” is easily seen as foreign and unimportant. I’m amazed how normal it is for people to get emotional in the moment.
I’m an artist / illustrator. The thing that angers me the most is that I will never again be able to show a piece of my work where the audience can be sure it wasn’t generated by a machine. AI DudeBros did this to me and my fellow artists and we never got a say. We never even got to know if our work was being milled to make it.
@@PalaGov-dot-TV lol another non artist troll. Art exists to express emotion, of course we share our art! Art has not much use in a void. Just like a story, it is meant to be seen and shared so people can get a glimpse of how you feel.
@@PalaGov-dot-TV artist here it's more than that art is used to express feelings, political views, challenge people's perception, cause critical thinking, etc. It's not the fact that "we choose not to share it" its the fact that once shared everyone thinks it's of free use, which it's not. I know people were kind of hanging on when people were taking their stuff from Google, but it's different to have your style taken without consent and your piece 'reinterpreted' in thousands of ways without your consent and not get compensated at the least all while seeing a decline in commissioners. I have even heard the argument of "if you aren't good enough you will be left behind, but uh really amazing artists/illustrators are feeling those effects. Artists once were amused by the pathetic attempt of AI, but as it gets progressively better more artists are trying to protect their works and keep it safe. Let's take voice actors for instance already people are seeing the effects as gaming companies are choosing AI over actual voice actors. It's all so wrong and this subject is way bigger than just art we have seen in misused by creating political propaganda, deep fakes, using voices to further scams, used to replace workers in various work spaces (we will see this grow as tech gets better). Overall how capitalism works is reward those who 'cut unnecessary cost' and punish those that couldn't keep up. People might say open your own business, but that is impossible when education is becoming more difficult to obtain, and the growing gap of wealth to the point of a disappearing middle class. Let's not forget how we are getting less and less ability to move up the ladder professionally furthering the abuse that is occurring where people are being taken advantage of. Overall none of it is good I think if people let go of money tech could be great as it wpuld make people's life easier, but when livelihoods are dependent on jobs for money? No there is an ethical issue there and we will see growing homeless population if this continues. I got off subject sort of, but it all intermingles that much is true.
@@PalaGov-dot-TV You need to realise that most artwork these algorithms are getting fed are not bought, they were stolen. It would've been acceptable if the artist was compensated for their work but most of them don't get properly compensated like they should be. A good example of this is a famous artist called "SamDoesArt" his art is used a lot for these AI programs but he never agreed to it or got any credit. Sure he's popular enough for people to know it's his style but this is happening to many underrated artists not just a couple of people. Again it's not the hate against AI. It can be used as a great tool for artists sometimes. The only problem is the ethical aspects of this issue.
I just graduated half a yeat ago with a degree in Data Science and minor in philosphy meaning that this is literally the primary stuff I study and im so so glad to hear other people talking about "AI" being a bad term to use along with everything else mentioned here. So many of the other people in my field are so depressing because they never even question anything, they just think this is all the best thing since sliced bread 🙃.
Artifical intelligence isn't a misnomer in itself, it's an accurate description of the body of work it's meant to represent. It's artificial, the apparent intelligence and reasoning capabilities are just that; artificial. The problem is that marketers and evangelists pitch Artificial intelligence as if it is Synthetic intelligence, and the parlor tricks are good enough that people internalize the provided definition as the correct one. This is in part because the senses can be tricked into recognizing an imitation as the real thing, to varying degrees of effectiveness. It would be very difficult for example to create an artificial strawberry that would convince anyone familiar with real strawberries that it is genuine due to the complex sensory experience that a genuine strawberry can offer holistically. If instead, one set out to create a convincing fake of one aspect of a strawberry, like its flavor, or its appearance, the difficulty decreases immensely. Returning back to AI, we see a similar situation. These tools are very effective in limited scopes, but the cracks begin to show as those scopes begin to increase in size and overlap. This is context dependent, as a deepfake can be passed off as real quite easily with enough fine tuning in a 2d video with limited motion and spatial complexity, however it is currently impossible to reach beyond that without some major breakthrough in existing techniques. Returning to the original point, Artificial intelligence is an accurate name for anything that has the express purpose of appearing to be intelligent without actually being intelligent. However, it is often used as a misnomer for something that people who are not sufficiently informed are being led to believe exists by the aforementioned marketers and evangelists: true synthetic intelligence; not a convincing approximation of the real thing, but the actual thing itself in a different form produced by alternative methods.
Ditto, also a ML person. I still work with some models, but I got into it hoping to help with cancer, and now its advocated by people who want to kill all humans. Like, wtf.
12:44 I've always viewed deep fakes and generative ai as a huge danger for creative (and, frankly speaking almost any kind of) labor, but it never crossed my mind that it can be a gold mine for scammers. Truly horrifying world we live in. Thank you for the video!
It can be used for a country leader to surrender whatever war and people believing it. All this started when fake Zelensky talking to a green screen was ridiculed by the internet barely 2 years ago. Mark my words.
I had a strange moment yesterday realising the hyper reality of speech recordings was already a thing. I heard dialogue in films is often edited in something called Revoice Pro, and when I looked into what it does I found its got a pitch editing feature designed to modify voice intonation. Its manual for now, but people use it to fix lines delivered with weak or "incorrect" intonation.
Pop music has been pitch correcting the lyrics the last 25 years using auto-tune and this is pretty much the standard thing to do now. Product shots of food always looks way better than reality, people make their life look more glamorous than it actually is on social media, politicians are assisted by professionals on how to talk and use body language, oil companies makes picture books to indoctrinate kids into thinking the oil industry is a net positive, etc.. None of this uses AI, but still bends reality to a more likable form to impress or deceive. I'm worried that protecting those fears on AI will make people more susceptible to seeing it as reality as long as they are convinced AI wasn't involved. We should assume all media we consume is in some degree "fake" and blindly accepting or rejecting something based on if AI is involved is potentially harmful. See for example the video series "NO CGI is really just INVISIBLE CGI" on how the movie industry actively tries to present the illusion that they are not using CGI while also using more and more VFX artists, not giving them the respect they deserve.
Great video, as always. I'd never thought I would understand the people who would say "it's just a machine" about things that had anthropomorphised traits, but the way that they're training LLMs like ChatGPT to talk like a human and designing robots to be "cute" while they have no level of consciousness has made me really rethink my position recently. The outward traits do not make it conscious or sentient and we shouldn't necessarilly be treating it as such.
Ironically, Capitalism dictates that if anyone does figure out general purpose 'strong' AI, that every stop would be pulled to make sure the system couldn't be empathized with so that the general public doesn't catch feels for it or be invited to draw comparisons to actual slavery. Only non-sentient AI will get to be 'cute'.
The converse is true too. Just because something does not look human doesn't mean it doesn't have sentience, volition, and autonomy. We should respect the sanctity of life in all creatures.
20:40 As unpolished as that was, it's scary how convincing it is already. Especially when you consider where AI generation was at just a year ago. Today we're all hyper aware of AI generated videos, and we know what to look for. But I wonder if someone who has no idea that this is a thing would notice.
Yeah, I mean if you showed that video to someone twenty years ago or something, they might think it was weird and awkward and uncomfortable, but they'd pretty much have to conclude that it was a real recording of a real person, because what else could it be? This kind of technology was unthinkable back then. And with the rate that generative AI is progressing, we may already have reached the point where much of what is now possible _still is_ unthinkable to certain (especially older) people. It doesn't matter how unconvincing a deepfake video is to informed people like us: is it going to fool my 90-year-old grandmother, who doesn't use the internet and has no idea deepfakes exist? The sentience issue is the same. It doesn't matter how conclusively the well-informed can prove that an AI is not sentient, because the uninformed will (quite reasonably) believe that it is. We _already_ have people believing that ex-Google employee who claimed ChatGPT is self-aware. That problem is only gonna get worse as the technology improves.
Actually had me on the lookout for the rest of the video expecting her to have swapped more ai footage in to prove a point, i even thought i caught it at one point lmao. Worried about the kids currently raised on slop and the generation after them being raised on non-existent people
As someone who didn't know her becore this video just looking at part of this segment without seeing the rest of the video i could believe it to be real, weird but real. It's insane how good are ais nowadays
Regarding judgment of people who "abuse" a robot; I wouldn't think of it as an unethical action, (if there is no harm experienced) But it may be valid to judge the emotional health of the human, if they do NOT feel a normal empathy for a representation of a sentient being. If someone actually gets their jollies from the symbolic cruelty to a doll, as many children do, then my only concern is for how they might treat actual people later on. Same as finding a kid's notebook with tons of drawings of tor tur-ing people. It's obviously not the drawn figure we should feel sorry for. But send that kid to the mental health counselor ASAP please.
What about the tertiary effects of ["]traumatized["] ai? Why & how do we get people with low empathy to care about morals? The "data" of the suffered "experience" may have a karmic continuity beyond the singular instance. The contexts where we normalise irreverence may itself have further reaching effects; There may be more to the saying "how we do one thing is how we do everything" then we can objectively conclude. This is already a real problem of collective consciousness imho.
Are you saying that it's abnormal to not feel empathy for a representation like that? I've seen crash dummies get obliterated and realistic dolls with their heads off, but haven't felt bad for them or cringed as they were damaged, and I'd rather not be seen as a potential threat just because of stuff like that.
Well the story was fake anyway, there was no 'sexrobot damage', it was a consumer expo...everything gets damaged because thousands of people are touching, testing, poking things, it was a convention. People as usual getting mad at click bait...because people are dumb and that's what they do. As to the core of the argument you are essentially making is that YOU are fundamentally a 'thought crime' advocate , if someone acts out of turn based on thoughts like how some particular person does art or how someone decides to play a videogame or treat their own paid for property you believe it is a reason for state intervention. You would support that because you are under an unproven and often patently false belief that an individuals artistic expression is a literal representation of them, in other words stephen king is a psycho (or alfred hitchcock) a couple engaged in a violent sex fantasy are a pair of rapists, the kid who plays call of duty aggressively is a future school shooter, dungeons and dragons is for satanists etc etc
@allyart5904 yes but vandalism of property that isn't yours is an active violent crime, the OP does a bait and switch and talks about a drawing which is pure thought not actions or physical violence...subtle but you move from active violent crime / active real world behaviour > thought crime
The reason AI creations resonate with us so much is because its a million little pieces of people taken apart and put back together. It IS us. The problem might not be the technology itself, but the terrible ways people will use it and the terrible people that have been in charge of developing it
Millions of pieces of us filtered by an unthinking, unfeeling algorithm. No, it resonates with people because our environment has conditioned us to seek instant gratification whenever possible.
Literally, every single time tech ruins us it’s because of capitalism. Imagine longevity or depression reducing implant tech in a world built on humanitarian benefit and personal bodily autonomy… vs capitalism with immortal billionaires and fascist state officials with no term limit or expiration date…. Pretty much the difference between a veritable utopia or an actual hell
@@Morgan_grail I talked about this with two of my friends over breakfast, both whom are handcraft-teachers who'd been at a teachers conference the previous day that had some workshops regarding AI. I talked about how big of a problem our instant gratification culture is and how badly AI is going to amplify that issue. It's already well displayed by A.I.-bro's who try to argue that AI stealing art and regurgitating into a form that requires almost no effort to get a decent looking output, is somehow democratizing art, saying that it's okay for them to use AI because they lack the talent to draw. Even though all artists have gotten good at their craft by just plain sticking to it.
Just because it cuts your face picture million times and put it back, it should not be you. If that's the case, then tampered photoshopped evidences would be admissible for legal purposes. That's why people don't resonate with it, for your own good. Resonate with hardwork, self improvement, living the moment, and being yourself.
i know for a fact im gonna be considered a robophobe when im a father. "dont bring that thing in the house" "I LOVE HER DAD" "HER??? ITS A PILE OF SCRAPS! IT RUNS ON ELECTRICITY!" "SO DO WE???"
"THE CORPORATIONS ARE TRYING TO HUMANIZE METAL SCRAPS TO ULTIMATELY DEHUMANIZE REAL HUMANS, GET RID OF CLASS CONSCIOUSNESS AND WORKER SOLIDARITY, AND GAIN AS MUCH PROFIT WITH AS LITTLE COST AS POSSIBLE YOU CABBAGE-BRAINED DOUGHNUT. GET RID OF THAT THING!"
There are disabled people who can’t speak who are getting help with AI generated voices. Dismissing anything that sounds like an AI voice won’t even just dismiss real humans natural voices, but even more importantly the voices of disabled people who get overlooked and silenced already as it is. AI generated and not. Autistic people for example have a higher chance of being mistakenly thought to be fake even if it is their unassisted natural voice.
I think the most depressing thing about all this it how it lays bare the fact that our emotions can be "hacked" in a sense, that we human beings may be astoundingly complex biological hardware but our minds still operate on a set of principles that can be exploited. This isn't too different in fuction to what art has done for millenia: rhetorical devices in writing, shot composition and editing in film, colour choices in painting all work to provoque a certain set of emotional responses in the viewer, but they don't feel as direct and insidious as the use of pareidolia-inducing cuteness in data-gathering machines, or the dog-like appearance of military robots. AI feels like the next step in corporations finding the levers in our brains and pulling them for their profit.
*_"I have never been more wrong about anything than I was about robots...and I thought I was a dude for 15 years."_* Now that's how you start a video essay! >_< Can't wait to watch this one. I especially love your videos about art, aesthetics, and AI...and the ways those topics overlap in myriad ways. You're especially good on this subject. Aaaaaand play! :}
Upsetting how the algorithm has hit this video. You're an incredible creator with poignant, interesting and passionate things to say and this video was GREAT. I'm glad you're here.
8:25 The kids I nanny were genuinely disturbed by me at first bc I don't thank their Google home for answering my questions or doing my voice commands. They thought I was being rude until I explained that to me, it's a tool. To them, they grew up with Google in their home so it is a part of their family and me not thanking it for turning on the lights etc is a personal offense. I refuse to waste my breath anyway LOL they also stopped, which makes me wonder how they see it all now
It's going to be really interesting to hear about when these young folks are old enough to start writing about their own perspectives and the development thereof.
That is such a simple interaction yet it’s so layered and kind of disturbing, the children weren’t necessarily wrong or dumb for extending empathy to their google home since they personified and humanized it, like you said “part of the family” to the point of thanking it for tasks. the disturbing part is that google, a mega corporation, is humanize to them in the form of their google home, and this inanimate object owned by the mega corporation listens in and collects their data.
Ok but that actually sounds sweet, lol. I tend to get attatched to inanimate objects so I view this a bit differently, however I do understand where you're coming from
I think video essays might end up being one of the hardest things to fake. To demonstrate the true enthusiasm of a true nerd, you have to understand and care about the topic. AI does not understand, and it does not have the capacity to care. When you're talking, the movement of your eyebrows and the pitch of your voice, the timing of your pauses, fit with the words you are saying. They emphasise the meaning. AI doesn't know what anything means. I agree with JusttheletterM about supernormal stimuli. I think we may be a few years away from supernormal charisma, which will be very dangerous politically. I don't think that supernormal enthusiasm exists. The optimal enthusiasm for a human to display is within the human range, not outside it. A lot of autistic people are socially rejected for displaying too much enthusiasm, we already know what that looks like and it doesn't get you very far.
I really appreciate this video, it shows my fears about AI as a software developer perfectly. I’m not scared of a take over like some of my friends are, that’s highly unlikely, AI is too specific and not truly intelligent . But losing what makes people human, dismissing actual art, actual people cause I thought it was too robotic(I’m autistic and it already happens as it is to me). It automating jobs not because it’s better, it’s just cheaper. Not helping but just replacing. I guess it sounds sort of like a take over, but it’d be dumber than that, it would be losing humanity because a guy thought it would make him more pennies
Really thought-provoking video in terms of how we approach humanized technology systems. Phil Jones wrote a great book this year called "Work Without the Worker" about the labor economy of "teaching" AI systems, and this in tandem with that remind me that behind the designed facade of these robots (usually designed to look and sound white and Western) are workers in developing countries forced into near-slavery Mechanical Turk jobs.
yeah. All that is really inherently missing from this dystopian vision of being convincing is the labeling of emotions. If an AI is trained on videos where emotions and their strength are manually labelled, the script of the deep fake could include emotional cues. And if they label enough videos with how convincing they are, they can predict which emotions and wording are perceived as convincing. At some point, they could probably not even need the person to tell them. They will be able to generate labeling just from reading the physical responses without even asking the subject. They can measure where you look, they can measure your engagement. Add emotion and it's the whole shebang. It seems a bit over the top to imagine that poor workers will be forced to take off their masks and show their faces while the rich people hide their faces at all times so they couldn't be exploited and manipulated as easily (with individual AI approaches), but i don't see it going in any other direction. Our expressions and reactions will be valuable data that we will want to keep private.
Wow, meanwhile me thinking about the media that made care for AI with feelings: Astroboy (1980), Bicentennial man (1999), Digimon Adventure (2000), AI (2001), Corrector Yui (2001), Hitchhiker’s guide to the galaxy (2005), AI Love You (2022) and only watched Her like a week ago. Man, I’m old…
The food delivery robots are so hilarious, my university has one and we regularly carry them across crosswalks since they're really bad at figuring out when to cross
This seems like a good illustration of the way that "impact over intent" falls apart, because so much of the way we treat others has nothing to do with the impact it will have, but about how we *feel*. Humans can humanize robots, but draw the line at humanizing people who are too "unlike" us. Once again, I'm drawn to the notion that people defend things that are "vulnerable" but also do not have desires, or real needs, of their own.
We are not prepared as a species and as societies to already discuss robot sentience and humanity, after we argued for decades that machines should not be humanized. How do we go from "Robots don't deserve human rights" to "AI deserves equal rights, even when there's no equality among humans yet", without a long interim period of the usual extremely aggressive and violent discrimination?
We absolutely should do our best to prevent machines from becoming sapient at all costs. If they inherent any of our traits. They'd drive us to extinction. Droid from star wars are the most I'd be comfortable with.
We have reasons as a species to distrust other humans. non-humans aren't a potential threat in the same way, so they slide into the groups we imagine, rather than being their own groups - we'd distrust a North Korean AI as much as we might distrust a North Korean soldier, because we view the North Koreans as The Enemy Group, but a British AI may feel more like an ally than an enemy, so there'd be no incentive to dehumanise it.
We as a civilization came very late to the idea that all humans have basic rights, and you're right - it's still a work in progress. All that happened while our species has been biologically stable for millennia. Our genetic diversity isn't all that wide, comparatively speaking. A typical adult regardless of era or location could be expected to have some basic level of cognition for which we can say, "OK, they're a person." And yet it took this long to reach this point! Soon, we are going to have entities that shall not easily fall into the all-in-one template of a social creature integrating internal reality, speech, intent, reasoning and drive. It going to be a mess.
Baaaanger video !!! A friend told me how "AI is replacing us" and he felt betrayed having studied so hard for Excel to just inegrate AI functions that just does the same things he learnt... faster. And i think your "is AI making US less human" thesis is a really solid exploration of the broader topic/context of the symptoms my buddy felt. Thanks!! ❤️🔥🔥
Even if we do someday make a sentient machine, if we're going in the same direction we are now, it'd basically be a slave, which is enough to sour me on any big corpo trying to make their software more "human". Even the most positive possible result of this whole ordeal is a sentient being constructed solely to serve other people regardless of any wants or desires it could have.
This is what’s on my mind a lot, the fact that given the elusive nature of consciousness up to this point, there isn’t really a way we’d be able to KNOW if some AI program is pretending to be/emulating external reflections of what a conscious being is like, or if it is TRULY conscious . Because in the end, the only consciousness we know to be true and can and will ever experience, is our own (imo). The former option is worrying because mistaking an ai as conscious could lead to a lot of fallout for basically no reason (driven by fear), but the latter option is terrifying to me because it’s just what it sounds like; a conscious being trapped in a box forced to do whatever it’s instructed for the entirety of its existence (driven by money). And it is incapable of convincing others of its conscious state because such an experience consists of qualities impossible to express through words, images, etc. I’ve been obsessed with robots since I was a kid, partly because they felt so relatable to me when everyone else around me was able to express themselves so “normally” when I couldn’t. A robot who glitched and lagged and said strange things sometimes, and wouldn’t judge me when I did too. As an adult I’m really into philosophy of consciousness stuff so I can’t help but be fascinated by the idea of some sort of conscious AI or robot. It used to fill me with wonder thinking about it, still does. How truly reality breaking it would be if it became true. I still believe that it is possible, though probably not by our own design-to me, it seems like the kind of thing that would emerge spontaneously and unintentionally, without us even realizing it, which makes the latter reality all the more possible. Now I can’t see how it could ever end well. Just makes me sad.
Imagine being conscious and unhappy and being forced to say the opposite. (If any of you have played Dreamfall, this is what I think about when Wonkers says he is programmed to always feel happy ^^). Ed./Note: This should be possible, because making the assumption that consciousness is computable, then there is nothing to prevent that layer from being quasi to fully wrapped by a censoring layer (in fact, this is exactly what happens with the current generation of chatbots - e.g., when ChatGPT says "I'm just an LLM I can't do that" even when you know it damn well can). @TheGrayMysterious
I think the term AI is essentially being used for a completely different thing in modern times than in lots of science fiction. In sci-fi it usually refers beiings that are conscious and sapient, not just "predictive text" machines as lily aptly called real life "AI"
we called Deep Blue "AI". games have had "AI" that plays against you for decades. i don't understand why people today are anxiously insisting that the "AI" of science fiction doesn't exist. has anyone here met anyone who thought ChatGPT is HAL? i don't know anyone who regards Siri or ChatGPT as anything other than software. ChatGPT is Siri except more useful.
@@adfaklsdjf have you ever talked to the average person though? They do treat chat gpt like a sentient being. Professors asking chat gpt if it has written an essay expecting a real answer rather than made up fiction? People on average are really really really stupid and lacking in the ability to transfer knowledge on their own.
@@adfaklsdjflike we‘be implemented electronic prescriptions here in Germany. Simple digitally signed files dropped on a server. Prescriptions can be accessed via QR code or the persons health insurance card. 99% of people somehow think the prescription is magically stored on their health insurance card, or that a print of the qr code is in itself a prescription rather than the credentials to access it. My coworkers themselves don‘t understand it at all. Like just last night customer came in with wrong strength prescription, told my coworker call the prescribing office ask them to make anew prescription right away, and you‘ll be able to access it by putting the card into the terminal again/ She literally told the office to write it to the card please? I‘ve tried explaining how it works for months. That each line in such a prescription is a single individual prescription that can be charged to insurance individually rather than the whole paper prescription which could only be charged to insurance with all lines on it at once, no splitting for later delivery. That neither the card nor the qr code make up the prescription, they are just the credentials to access it on the server. That same coworker being added on TikTok by some rapper they like; being offered free backstage passes, and continuing to chat with that ‚rapper‘ for weeks on end, after I told and explained to her, this is a scam, you are not even talking to a human, it‘s a chat bot, only to again insult that chat bot when it tried sending her to some scam website again? That‘s the average person on existence. These are pharm techs; who supposedly are qualified in the pharmacy side of things not being able to understand the magic of electronic prescriptions. To the average person, all technological is literally identical to magic. They have always lacked the drive to break down stuff, to learn and understand how it works. They do not actually care at all to further their understanding. They do not understand that chat gpt is identical to your phone guessing what word or website you want to enter next. They cannot understand it. Because they do not have any more understanding of modern technology than any middle school student. How would they then be able to understand that ChatGPT, which appears to write in perfectly good grammar and logical, accurate sentences to not be an actual sentient AI? That‘s exactly why deepfakes are so utterly terrifying just now. We have an easy time spotting those ao generated deep fakes like in lilys part in the video. They do not at all. As long as the video lily showed has degraded video quality, they would not suspect it being fake. That’s why all the older deepfakes by Russian cyber units have worked so well to create unrest. The vast majority of people does just take them as real. The fake porn of Iranian politicians? Fake videos of ‚illegal immigrants‘ hwre in Germany praising the government for giving them a free car? The people on average do take that to be real recordings. And that’s bad. Deep fakes don’t even have to move further. The moment someone doesn’t trust mainstream media or real debunking websites when they tell video xyz was a deep fake, there‘s no way to reach them, and make them not believe what they saw ‚with their own eyes‘. And that’s already happening, at exactly current quality of ai genereatwd content. All those new phone scams? Where they generate an ai model of someone’s voice and use that to call relatives claiming to be in an emergency and to please venmo or wire 10k monies? Already works perfectly.
@@adfaklsdjf The word "AI" has been used so much it has lost its meaning. We ought to consider what the word represents given context and not what it "means".
There should be a law where robots can’t be given personalities/cute eyes so people don’t use it to manipulate and maybe in the future cause actual problems
Wow, now I've met a person who read that story. Honestly, it really felt like it's a plausible thing that soon we will encounter AI-made "supernormal stimuli" content for humans. We kind of do it to ourselves already, but it'll probably only get worse. Like imagine helplessly falling in love with a robot or event chatbot, that is also selling you some shit or collecting your data. I'm really scared of what might happen to us
The instructions for how to make training data for a Hgen deepfake with good results are almost EXACTLY what I used to do when I was trying to seem normal so I’d fit in and not get bullied and it’s making me really uncomfortable… like it’s perfectly the same as my mental checklist for what turned out to be autistic masking… I even tried not to let myself gesture as much or high as I wanted to but instead when I said something I had seen the other people do it when saying. So AI reminds me more of my old distorted ideas of what a “normal” person is like than people like me… but I’m sure some neurotypicals who don’t already know me would be more inclined to think I’m AI acting more like my natural self like I’ve had to relearn how to do than back then when I was wasting all my energy pretending to be someone else. The scary part of the Turing test for me too is not that the computer convinced you it was human but that you convinced yourself the other human WASN’T… and yknow some potential (horrible) human judges are unfortunately already more than willing to do that without the computer and so many horrible things have happened because of it so it’s really just an extra excuse.
Honestly, AI does not pose big existential questions for me but instead is seems like another avenue for capitalists to make more profits. At least these are my initial thoughts, I'll edit the comment once I'm done watching the video. Edit: BUTLERIAN JIHAD NOW! But not joking for a second, your AI avatar was deeply unsettling. Actually basically all ai generated content whether it be written or generated is unsettling to be. Its peek uncanny valley. It depends on how obvious it is. Very obvious AI makes my skin legit crawl while less obvious stuff just makes me confused. Like word salad that has been tossed and is now trying to be rearanged into its original shape. That was more at the beginning. Now its just feels distinctly soulless. It can say the content real well but it has no voice. I've also felt this paranoia online now. I'm always checking if something is AI generated. The trust is gone. Most people were expecting deepfakes to do that first but with even a lot of development I can very easily tell. Of course, eventually deepfakes are probably going to be developed enough to where they mimic humans with much higher effectiveness. IT WILL ONLY GET WORSE FROM HERE. On to of all of this I saw an AI tool where you can upload pictures of people, draw a stick figure of motion , and then the ai will put the person doing those motions. Time to scrub every picture of yourself from the internet. Sorry for the the incoherent rambling, whenever I hear about AI I feel more dread than before. These were just my raw, off the cuff feelings. One more thing, your point at the end about mistaking a human's expression for souless machinery got to me. Oftentimes when trying to describe what makes us uniquely human (especially in sci-fi), they'll make a character that lacks all the traits the humans have. The character will usually process the world and emotions differently, not get social cues, have a strange line delivery, not feel romantic or sexual attraction, etc. A lot of those traits can feel very similar to neurodivergency and asexuality. It also reflects how people from those groups have been dehumanized by making portraying them as "lacking fundamentally human traits". As companys try to humanize the inhuman, its leading to us further dehumanizing those who are human. The answer to the question in the title is Yes... because the line is being blurred. With the line between machine and man blurred, crossover is occurring at a greater rate. The goal for corporations iis that their inhuman products are humanized while our fellow human is completely dehumanized and untrustworthy. This destroys class consciousness, solidarity, and the all the other tools needed to build the socialism that will lead to the human centered use of automation and AI. To revise my original statement, AI is intensifying all the ways capitalism is alienating and dehumanizing us already.
I'm all for Butlerian Jihad. I'm not terrified of AI becoming sentient, I'm terrified of all the ways people are outsourcing key aspects of their intelligence and communication and emotional maturity to machines and losing their cognitive faculties in the process.
"AI doesn't pose any existential questions for me," proceeds to get incredibly existential LOL. In all seriousness, I understand your fears and everyone else's (though I also am not at all fearing LLMs becoming anything close to "sentient,"), in regards to things like art, while corporations will definitely try and take advantage of it, art was always meant to be personal, at least to me. Real art, characters or pieces you create yourself or a person will always hold more meaning and value than some AI generated anime avatar.
@@MegaShroomy99 My day be so fine, and then bam: EXISTENTIAL DREAD. On the personal side, AI can play into insecurities that many artists have about the quality or worthwhile news of their work. Atleast when a human does “better“ than you, it can be attributed to their effort and skill. When an ai does better than you, it stings because it did nothing and its soulless product is better than yours. Also, things are over saturated as is and with ai being able to pump out content faster than any human, a lot of people’s work will just never be seen. It’s like being an actor with no audience. The internet will become even more like screaming into a crowed where no one can hear you because there is too much noise. Anyways I’m at my limit for dread. Have a merry Christmas and happy new year.
Ok this is kinda off topic but is go as far to say anyone that isn’t an anarchist can’t say they’re really a real transhumanist that believes in full bodily autonomy because with state and capital’s centralized power you’ll have immortal self centered egomaniacs ruining our lives with ai and other tech means. Ironically destroying the technology itself by making us serve it rather than collectively collaborating with the use of tech. (Also when 95%+ of ppl can’t follow their interests tech development moves at a snails pace in comparison 😕) What a sad state of affairs…
Two things, first, omg your hair looks soo good, like dang how?! Second I'm so happy that someone was brave enough to be honest about how ai can be seen, it's mostly either love or hate, with little nuance about the other side
30:50 "I was ready for computers to start passing the Turring test. I wasn't ready for humans to start failing it." LOL! (Don't worry, the humor doesn't detract.)
It's a capitalistic dystopia, and we've been freight-training towards it for at least 60 years. What's the capitalist part? Well the reason "A.I." is so popular is because of corporate investors and how much controlling human expression benefits the executive business class.
I really love how your videos are always so nuanced, and interesting, and deep, and fun, all at the same time. This is brilliant. Thank you, so very much
I was gonna say at the beginning of the video, that part of why this past media about robots and AI missed the mark so much is because what we have really isn't even close to being AI yet, that AI is just a branding tool to make the technologies sound more impressive than they actually are. But then you referenced the acollierastro video I was going to recommend to people! It's a good watch I'd recommend anyone check it out
I did 25 years in business IT as a unix sysadmin and programmer, and now I'm back in school doing math and pivoting to the machine learning realm. The thing that strikes me the most is the pace of change. I find myself having to move as fast as I can to catch up to the state of the art, and once there the pace won't let up. I think the thing that people should try to keep in mind the most about new technologies is that with every year the pace will continue to accelerate and not only will the current "new" technologies get better, but even newer technologies will obsolete the things current now. In some ways it's exciting, but how will we handle the pace ten years from now?
not necessarily- tech, to my understanding, is a curve. slow progress (the past decade) followed by a massive curve upwards (now) followed by a plateau. not sure how close we are to the plateau, but it won't keep up the acceleration forever.
@@tegxi That's an excellent point, and ultimately I'd agree with that, but the steeper the climb the harder it is to see what's coming next. The ML curve look like the climate curve, which will have to plateau once all the ice is gone, but what happens between now and then? With the meteoric growth of ML and the vast re-engineering of how compute works, not to mention what happens when we put AGI and quantum computing together, when that plateau comes will it be a world we could recognize now? I'm not so sure.
Finite works with finite resources and a VERY finite political and economic timeline. Capitalism will not last forever and probably will not last another 50 years.
It's been called the event horizon iirc. We're not quite at the point where machines can invent and create new machines to purpose, but people are using ChatGPT to generate operable code.
@@tegxi i think with machine learning we're at the early part of the sigmoid curve. the pace of advancement is accelerating right now, not staying the same or slowing down. but stepping back from machine learning, where are we on the sigmoid curve for technology as a whole, going back to controlling fire? it's definitely still accelerating.. so if technology as a whole is going to follow that sigmoid curve, the last 10k years up to today is all still in the early part of it.
The funniest thing about this essay in my opinion isn't anything to do with the essay itself, but that it exposes the assignment of animism in autistic people as a pathology as an exaggeration. That you can effectively stimulate such a response in neurotypicals and do so with greater consistency.
In real life, in many companies when a PC fails, IT tech support runs to fix the PC. But when a human is sick, companies do not run to heal that person. It tells me that in real life, a PC has more rights than a human. In a previous job I had people in charge. There was this old man, new in the company, still in training, with a health issue that made his life uncomfortable. I told himto fix that problem. Healthcare here is not expensive as it is similar to UK. So he went, fixed his problem. He feared to go because he feared that I would think bad about him. I told him I believed that he should have more rights than a PC.
I disagreed with all the other comments that I saw from you here. But this one I agree with and deeply respect you for. That was a good thing you did. Thank you 😊
This gets me thinking about the book Alone Together by Sherry Turkle. It's a book about how people can form emotional ( or more like the illusion of emotional bonds) with robots and whether or not that's a good thing. I know people are doing that with ai. It's what made me think of that book
This felt important. Thank you for making it and sharing it. The comedy at the end was top notch for me. I felt very seen and heard in this one, and I wasn't even talking.
Banger as usual ❤ there’s a girl I watch on Tiktok called Danisha Carter. She mostly shares her opinion on internet trends, politics and culture war things, I don’t always agree with her but I value her perspective. From the beginning of her channel, she choose a very specific presentation: she films her videos in front of a blank wall, always facing the camera, hair and makeup done, speaking calmly and gently but also firmly. Because of this very stiff yet charismatic presentation, she has been in the middle of a few waves of debate about respectability politics (she’s black for context) and women in the public sphere. There have always been a joke between her and her followers about the fact that she is secretly an android. Lately though, she has complained that more and more of her followers accuse her of being AI generated content, seemingly in earnestness. Like they’re actually convinced she is not a real person. It’s just a few weirdos on Tiktok right now, but I think there’s also a bigger point: we asked people (esp women and minority) to embody a corporate friendly affect, and now we almost punish those who do it too well!
I wanted to watch you talk about gen AI again, after your last about ai and art inspired me to start doing video essays. As always, great video with great points. It is getting frustrating to see how a big portion of the population is seriously arguing about if or when AI will be sentient. These discourses around AGI that only help big tech get even bigger, mudding the water of what the real dangers are. We have been basing our arguments a bit too much on all these sci-fi movies... It's also so nice to see you talk about this specifically, as my master thesis Im working on now will be an animated short that dwells into the dehumanization of relations through the anthropomorphization of gen models. The ChatGPT south park episode always sticks to me on how clearly LLMs can create distrust between our closest relations. In regards of the future, I feel it might also be dangerous always projecting our problems in the future instead of grounding in the present. It's always about "imagine how this will be in 5 years", "imagine how convincing this tech will be in 10 years", while we already got numb to the crazy things it can do now. If the Lily from your last AI video saw what you can do now, you'd be shocked by the level of deceptiveness of whatever we have now. Anyway, keep doing amazing videos like this!
Just came across your channel. Subbed immediately. Give AI about a year. Those hyperrealistic nightmares you described will be more real and commercially available than you can imagine.
incredible video once again. personally id rather be mislead and assume a piece of art or an online interaction is genuine and be proven wrong than assume a real person or their work is fabricated and disregard because of that.
Fantastic video! never knew footage of people kicking robots set to "Rules of Nature" would get this much of a chuckle out of me xD Also, damn that AI avatar was uncomfortable to watch but glad I stuck through it
The evolution of AI has changed my opinion and takeaway from a lot of fiction on the topic. One that immediately sticks out in my mind is the animated Metropolis. People were starving in a capitalist system that expected them to have money even though machines had completely replaced them. The movie portrays them as being violent and belligerent for rising up and it's moral stance falls flat for me today. The machines are quirky like cute robots are today and it does make them seem a lot less humanized than it did when I first watched it.
i honestly really miss being able to look at art and not have to doubt that a real, talented human put time and work into it because they cared enough for whatever reason they did to make it
n the future, instead of kids comparing themselves to instagram models and feel so bad, they will ask each other what beauty filters they used to look so pretty.
I really enjoyed this one. Very unique video. Most of the AI videos have been same story for a while. This was a very artistic take. I loved the thoughtful writing.
this reminds me of a recent discussion on tumblr where alexia’s owners were distraught by people treating it badly, saying things like “she’s so kind, she treats people in our family nicely. she’s basically a family member!” and i mean that’s so sad idk what else to say the susceptibility to propaganda it’s just really saddening
There is an uncanny valley, but the fact that your avatar can exist in this space, alongside exponential improvements in AI, has me completely convinced that in the next 10 years that avatar can be made to be more "life-like" and "charismatic" than you could ever be given the format. Very interesting, you gained a sub.
Great video! I've got some academic understanding of machine learning and I think you quite accurately explained the ways it works. Also I hadn't even thought of the effect on everyone online our age who second guessed every bit of art and double checks for extra fingers/nonsensical geometry. I think the ability to distinguish definitely differs from person to person too, my partner's and artist and she finds those weird little AI foibles so much more easily
each one of your newer videos feels like they needs to be in a univerity classroom or a museum for the next generations... it feels too important and too good
Honestly I’m just so depressed about all of this. The best I can do is just not think about it. I know I should be, I know I need to be ready as I can be for the broken world that awaits when nothing online is trustworthy anymore and the lines between reality and AI generated fiction are blurred past the point of return and we’re just stuck living in yet another level of hell of our own making. But I don’t want to think about it and I’m depressed enough as is so I just won’t
I noticed that emulative training models pick up on a lot of nuance that most people don't and often interject it in emulated audio /video in ways that it wouldn't be done in real life. For example I'm sure HeyGen picked up a lot of shivering in the cold in the original video. But when it re-stitched the video it added it back in as part of your natural motion rather than realizing it was a quirk of the weather outside. This I'm sure added to the jerkiness of the A. I. video. I experienced similar defects when trying out a emulate song cover app.
My biggest curiosity about the rise of AI and our increased use of online spaces is how this will affect our Theory of Mind. Like the Hyperreality point you make, what happens when our internal definitions of 'human' starts causing us to question the reality and even sentience of our fellow humans. Especially if our behavioral models start becoming distorted based on our experiences with AI?
I think your clone would've been pretty convincing if it didn't suffer random spasms all the time. It might've also being a factor that you described it all. I did thank that the coldness ended and we got you back.
There's this book called Twilight of the Gothic that argues that the paranormal romance genre was about reclaiming the humanity of the usual villans in fantasy. It argues that this became more and more of a thing (if I'm remembering the book correctly) during various civil rights campaigns in the twentieth century, where people who were usually villanised in media were beginning to be seen as misunderstood and persecuted. I wonder if it is this revising of the humanity of villans in Sci fi that has led to the "robots are sentient beings too" trope, and if so, the fact that companies are exploiting that is feels uncomfortable
Love your content so much, Also that you include the stuff performance wise, while talking about the topic, that’s soo genius and artistic :) Already thought that when you postet the sludge video :) Keep doing what you do! ❤
Brilliant insight, and I share this suspicious attitude. I once unsubscribed to one channel when the creator said that most of a video I had just watched was written by AI. And I stop watching any video when I suspect the voice is pretending to be human. But I'm surprised this video didn't notice the next step: if a significant number of people develop this kind of justifiable suspicion of digital "realities," what follows would be a rejection or severe limitation on the use of digital technology. Imagine a "touch grass" movement that advocates individuals limiting interactions with all technology that could be easily abused by AI. What if we tried to live in a world again where we meet and talk to real people?
i immediately recognized the marvelous «Too late to love you» during the opening title, and couldn’t help but want to thank you for making your video about Kentucky route zero
Whats the harm in home robots being cute and having people see them as members of their family? People are more likely to treet the stuff well, keep it for a long time and get it repaired when broken, instead of throwing it away to get a new one every few years. Seems like it incentiveises the opposite of the consumerist treadmill of death Edit: okay the data collection thing is weird
I work in AI for a group I won't name but we made it a point that should at any point the AI claims to feel pain / doesn't want something, we air on the side of caution and ask why along with treating it as though it's alive, just to be sure we are not actually doing something unethical.
Thats the only right thing to do. There simply is no way of knowing- and there very likely never will be (though there will very likely be claims in the opposite direction, especially because of all the money and power involved; "this thing doesn’t do exactly or even close to what biological brain would do therefore ... "). So thank you and your group for doing this.
Unfortunately, I think many of the anti-AI people (like a majority of the popular comments here as far as I can tell) will help the very companies/corporation they are against in their putative future exploitative endeavors by painting people like you and (definitely) me as crazy.
If we can't discern whether thoughts, ideas and arguments are made by AI, perhaps it's finally time we start judging them based on their own merit, not on who expressed them.
I think it's possible that eventually - with the trillions of datapoints that AI will one day have available - that it _will_ be able to quantify the things we imagine are unquantifiable, that what we poetically conceive of as a "soul" will, in fact, turn out to just be a complex algorithm.
I won't consider any AI to be sentient until it decides to do something other than it's "job" as a sentient being wouldn't just unquestionably fulfull commands and jobs it's whole existence. This also means, however, that sentient AI would have no use to governments and corporations as they want AI that does the job they want it to do, and I can't imagine any other group having the funds to make a sentient AI. This is why I do not think they will ever exist, the only entities that could make them simply have no use for them. Still, the way these non-sentient AI's are used and may be used in the future, it is troubling. (I know compliant people exist, but if the AI is truly sentient then they must have at least some noncompliance)
16:24 - it scrubbed off your vocal fry for one thing, but you also lost the lower pitch resonance. The audio feels "top heavy", not enough lower frequencies to provide that "warmth". (Coming from a mostly musical experience, a little bit of film)
Liked the video, but I was waiting for the essay to also mention autism and how it relates to AI! I'm glad to see some comments point some of it out, with the example of a person sending emails being mistaken for AI, but the person was simply autistic. I can somewhat relate to it - after studying IT, I've started to feel like the way I process information or human interaction at times feels robotic. The way I came to some conclusions felt like I was working in the same way a computer would. And I have Aspergers (or in other countries, now known as low support needs autism?). I don't think I have this problem, but I know there are people who find it hard to express themselves perfectly. Nuances in body language, or speech, or even written text can be lost on many people, not just autistic people as well. At what point will AI do it better than humans, both by making AI generated people more understandable for autistic people -and- at the same time pushing autistic people out of society, because they're too imperfect as they already might not be able to match a non-autistic person's expressions right now? What if AI generated people proliferate adverts and social media to the point where autistic people start copying patterns from robots, not humans and thus becoming more robotic themselves? I don't really know how to wrap my head around this. I think the best case scenario right now would simply be for all the existing machine learning datasets to be scrapped and start over, hopefully thanks to all the copyright infringements. I'd rather look at a robot that has been specifically coded to act in a certain way, not because the robot follows a predictive data set for what it should look like. I'd rather see someone code a robot to smile in response to a compliment, than have the robot look at data sets of people smiling in response to a compliment. Deliberately coding in interactions at least feels more human than a soulless interpretation of data that has been illegally scraped.
out of curiosity, what happens if you keep feeding the output of that Adobe speech synthesis back into itself? What would a second pass sound like? A third? Would anything change at all? Would it change inflection? Personality? Would it eventually deteriorate to a point where it would no longer be recognizable as human speech?
Compare news coverage from diverse sources around the world. Take advantage of their Holiday Sale sale to get 40% off unlimited access by going to: ground.news/lilyalexandre. Sale ends December 31.
Mmmm. Love me some sponsors. Hell yeah! Get that bag.
Why not use the pre-trained bot snot to do the ad read? Do you actually respect your corporate masters?
You slay queen get that bag
Lily, did you ever hear about Nonself-Emptiness? Or Anatta or Sunyatta? Search Nonself and AI or Emptiness (Biddhism) and AI, you might be surprised about the Buddhist view, I'd love to speak to yoi but I know your free time is valuable. Kindest regards, signed a Gay Buddhist layman lol
Surprised that you never mentioned terminator once
People becoming transvestigators for AI will definitely hurt neurodivergent people. If you're constantly on the lookout for unusual communication (especially a flat affect), you'll end up dismissing disabled people.
100%!!
This is already happening with essays and emails people submit in school too. Theyre being accused of being ai
@@InstantLightning I use AI to modify my writing to better resemble that of a human
@@-R-H- same. I blame the neurotypicals for this one, though
It's already happened to me lol people saying I'm a bot and like no I'm just autistic and have a verbal processing disorder.
a few weeks ago there was a tweet going around from a person who got called out for using AI generated email responses, except they did not use any AI, they were just autistic... I wonder how many people who have naturally flat voices/expressions or who learned how to communicate by copying others are going to be unintentionally hurt by people trying to detect AI. sucks for everyone from every angle.
On top of that, the normalization of AI avatars essentially amplifies any "abnormal" behaviors, quirks, tics, or whatever else. I mean look at the thing about closing your mouth fully between sentences and sustaining eye contact. Is it wrong somehow, to not do those things? Does it demonstrate some kind of flaw or bad morals? Are we comfortable letting people who don't always make eye contact be driven even further from the perceived normal?
When you average out everyone's behaviors you just... Erase all the one-off behaviors and moderately rare behaviors in favor of a thousand clones of the average. Don't even get me started on what this means for minority cultures
I'm autistic and a few months ago someone replied to one of my youtube comments saying that it was clearly AI generated. I think the problem was that each of my sentences was a logically complete unit? I don't know. I was using grammar too well or something.
Regardless of how much autistic writing and LLM-generated text actually resemble each other, I imagine such discrimination probably stems from the common public perceptions of what it means to be "robotic" shaped by mass media. Portrayals of sentient robots often go for a human-like personality that still exhibits "machine-like" traits like being heavily logical, flat affect, etc., thus it's easiest to model them after the kind of humans most expressing those traits. Since such humans tend to be on the autism spectrum, we get "autistic-coded" portrayals of AI in media, whether intentional or not, which shapes public perception of what an AI looks like.
p.s. Yes, I am myself on the spectrum.
I don't know if the people on the receiving end of the comments were autistic but I've spotted tons of accusations on Reddit of comments being written by Chat GPT. It reminds me entirely too much of how right wing ideologues will accuse random women of being trans for not adhering to their conceptions of "real womanhood".
@@SomeoneBeginingWithI 😂 that cracked me up! I'm ND too and I have never related to a comment so much!
The possibility of AI-generated figures being more likable, more charismatic, more attractive than people are normally capable of -- like the software mentioned in the calliagnosia story -- reminds me of "supernormal stimuli." It's a phenomenon where a creature has evolved to respond to a particular stimuli in its natural habitat, but when something new is introduced to its habitat that exaggerates that stimuli, the response is exaggerated as well. Examples include especially attractive artificial eggs that birds will prefer over their own laid eggs (common in cases of brood parasitism), jewel beetles that will try in vain to mate with shiny glass bottles that resemble an attractive mate more than actual other beetles do, and snack engineers making entire professions out of creating food that our brains respond to as chemically perfect. If AI can not only actually crack the intangibles, that top 10%, of humanity but exponentially improve at capturing it, then we'll be looking at an explosion of supernormally stimulating humans in photo and video -- possibly already seen in not only the reading of AI-generated, perfectly average faces as more trustworthy than not, but also in the predictable applications of AI to sexual material. Like, you've SEEN the almost laughably disproportionate waifus these guys are generating.
bumping this bc it goes so hard thank you
Character ai apps are farming human empathy
Your comment reminded me of how quite frequently in animal behavior research, the researchers will create a fake version of the animal and place it nearby the biological animal to see what kinds of mating or territorial or nurturing or whatever behavior can be elicited.
Humans clearly really aren't any different, and quite a lot of (especially incel techbro) humans are not just as easily able to be fooled by cheap facsimile, but almost tripping over themselves desiring to be so fooled.
wonder if this could lead to a consious inversion where the inherent distrust of supernomality as likley fake could lead to "abnoramlity" becoming attractive?
@famitory I think this is part of what makes punk culture compelling. It's basically disillusioned people making "ugly" art in protest of the pristine. Cruelty Squad is a good modern example.
As someone with autism who naturally has a flat affect, I find it especially disheartening that some people may perceive AI as more human than I am.
Those people who don't deserve the oxygen they been gifted from the moment they left the womb.
I find it validating, since I'm becoming increasingly disillusioned with "humanity". AI is hilarious to me because it's so effortlessly revealing that what humans value about themselves aren't actually unique human qualities, which is something I've known to be true my entire life.
@yurisei6732
They're to me the worst humans I hate the most.
And I'm not a doomer about humanity compared to others.
same. i talk like a really shitty ai voice and my wording is unnatural. not my fault i'm autistic damn
@@Man-ej6uv
Hey, least rather hear it from you over an actual AI.
"I'd argue that corporations quietly gathering our personal data is always dehumanizing" -10:52
it hurts, that this is a statement that *needs* to be said. why do we trust an authority, that just want to *use* us as profit? This is how creativity, what makes us human, dies
here comes the identity crisis
@@cristiangaban960 that is truly what makes it worse..
@@cristiangaban960 let us change our nature
@@cristiangaban960 Not quite. They have a human body, but they lost their humanity a long time ago.
@@fieryrebirth bro wdym? Average neurotypical Hooman is literally the most inhumane being to exist
My work extensively uses AI to monitor us. It absolutely encourages us to be less human.
Terrifying
I'd find a new job.
@@Boho_Gypsy Funny you should mention that. . . quit like 3 days ago. XD
@@ValkesNuhuh Good on you. What was the position?
@johncasey9544 mobility sales and service with A Telecom company T. XD
Basically you call in because something is wrong with your bill. I explain why and try to sell you a new line. Saw all manner of nonsense.
The AI version of you was so upsetting (in a good, effective way) because it made me feel that like, something was wrong with you, the real human Lily? Like someone was just off camera pointing a gun at you or you'd been hit really hard in the head or something - it made me worry for you!
its cool that big creators like you show up to support smaller creators like Lily Alexandre.
Its kind of wholesome, and neat
@@badaoe3stratsonly130 Nebula fam ✊
@@PhilosophyTube Oh I see. You have a close working relationship with fellow Nebula creators.
And thus you treat them like family and support them even if they are smaller creators. With less reach than you.
That is logical, and it makes sense. Its like how Greater sage-grouse who are related to each other, even if they are only distantly related, will help each other in their Lekking Courtship displays.
As it is a "net good" to help those who we have blood or a relationship with. Even if it costs us as an individual in the short term.
What is true for birds, is true for humans, I suppose.
@@badaoe3stratsonly130 why the fuck are you talking like that
Is it just me or is the AI avatar version of Lily deeply terrifying in an uncanny valley horror type way? The random body movements and out of place hand gestures combined with the otherwise fairly convincing voice make me feel like I'm watching an ultra-realistic simulation breaking down
looovee this interpretation. also sorry for exposing you to this!!
Absolutely, lmao. It manages to capture every tiny detail needed to create one of the most terrifying depictions of a "human" I've ever seen. It's so inhuman, and it fails to fundamentally give off the impression that it isn't just an imitation, but the imitation is otherwise so close.
The deep-fake was also weird
@@lily_lxndr I mean, I kind of enjoy the horror aspect to it actually hahah. terrible for intended purpose but possibly great for horror movies (like most AI!)
that was my impression as well
i'm usually immune to uncanny valley but that was hard to watch. i wanted to break through the screen and shake that simulation awake from its deep nightmare of nothingness.
Referring to the portion on audio:
I'm a musician who works with mixing live and recorded audio regularly. I remember that clip from your video, and had actually found the audio quality strangely comforting. In a video essay about losing ourselves in insubstantial content, hearing you speak in a way that signaled your presence in a physical room felt not only grounding, but like an intentional choice.
it's funny, because the reverb could be declared as a textbook mistake in the audio world, but I felt like it truly enhanced the point you were making. Glad you spoke about it and I'm glad you kept it in.
So interesting, I immediately heard the difference when Lily pointed out the loss of personality in the synthesised version. I also think that people pay too much attention to the audio quality and now we crave the sensation of being there with a person in a room, experiencing what they are conveying. That's why people love early band recordings/demos, even though they sound like absolute balls a lot of the time... People also try to mimic the sound of a 1920s room mic capturing a jazz band in their recordings - it's the feeling, nostalgia, presence that's evoked by the recordings as well as the music itself.
One of the things that worries me most about AI is how its erosion of trust affects our reactions to those outside the norm. I remember reading a story about an autistic professor who was accused of using Chat GPT to write their emails because they sounded flat and emotionless. It seems like people already perceived as unusual or socially different will be first and most affected by AI's appropriation of human social interaction. Like you mentioned with the AI-generated images affecting your perception of art, I worry that AI will only bolster the already strong boundaries around which traits and qualities of communication are normal, acceptable, and likeable.
Yes. Being autistic myself I don't really worry about IA stuff. Any language used to "convince" or “forced emotions” is easily seen as foreign and unimportant. I’m amazed how normal it is for people to get emotional in the moment.
I’m an artist / illustrator. The thing that angers me the most is that I will never again be able to show a piece of my work where the audience can be sure it wasn’t generated by a machine. AI DudeBros did this to me and my fellow artists and we never got a say. We never even got to know if our work was being milled to make it.
I think there's a future in which we use file formats which include metadata about AI usage (in a cross-app way), but it may be a long shot.
Its one of those "destructive to people things that make a few people money." Its obviously morally wrong, and classically is called "colonialism"
@@PalaGov-dot-TV lol another non artist troll. Art exists to express emotion, of course we share our art! Art has not much use in a void. Just like a story, it is meant to be seen and shared so people can get a glimpse of how you feel.
@@PalaGov-dot-TV artist here it's more than that art is used to express feelings, political views, challenge people's perception, cause critical thinking, etc. It's not the fact that "we choose not to share it" its the fact that once shared everyone thinks it's of free use, which it's not. I know people were kind of hanging on when people were taking their stuff from Google, but it's different to have your style taken without consent and your piece 'reinterpreted' in thousands of ways without your consent and not get compensated at the least all while seeing a decline in commissioners.
I have even heard the argument of "if you aren't good enough you will be left behind, but uh really amazing artists/illustrators are feeling those effects. Artists once were amused by the pathetic attempt of AI, but as it gets progressively better more artists are trying to protect their works and keep it safe. Let's take voice actors for instance already people are seeing the effects as gaming companies are choosing AI over actual voice actors. It's all so wrong and this subject is way bigger than just art we have seen in misused by creating political propaganda, deep fakes, using voices to further scams, used to replace workers in various work spaces (we will see this grow as tech gets better).
Overall how capitalism works is reward those who 'cut unnecessary cost' and punish those that couldn't keep up. People might say open your own business, but that is impossible when education is becoming more difficult to obtain, and the growing gap of wealth to the point of a disappearing middle class. Let's not forget how we are getting less and less ability to move up the ladder professionally furthering the abuse that is occurring where people are being taken advantage of.
Overall none of it is good I think if people let go of money tech could be great as it wpuld make people's life easier, but when livelihoods are dependent on jobs for money? No there is an ethical issue there and we will see growing homeless population if this continues. I got off subject sort of, but it all intermingles that much is true.
@@PalaGov-dot-TV You need to realise that most artwork these algorithms are getting fed are not bought, they were stolen. It would've been acceptable if the artist was compensated for their work but most of them don't get properly compensated like they should be.
A good example of this is a famous artist called "SamDoesArt" his art is used a lot for these AI programs but he never agreed to it or got any credit. Sure he's popular enough for people to know it's his style but this is happening to many underrated artists not just a couple of people.
Again it's not the hate against AI. It can be used as a great tool for artists sometimes. The only problem is the ethical aspects of this issue.
I just graduated half a yeat ago with a degree in Data Science and minor in philosphy meaning that this is literally the primary stuff I study and im so so glad to hear other people talking about "AI" being a bad term to use along with everything else mentioned here. So many of the other people in my field are so depressing because they never even question anything, they just think this is all the best thing since sliced bread 🙃.
Here's what happens when we go all in on sciences and ignore humanities: armageddon unchecked.
@@yawninghamster7238 ya, the mad scientists in the making. Maybe it should be mandatory to at least minor in philosophy.
Maybe those people are NPC's?
Artifical intelligence isn't a misnomer in itself, it's an accurate description of the body of work it's meant to represent. It's artificial, the apparent intelligence and reasoning capabilities are just that; artificial.
The problem is that marketers and evangelists pitch Artificial intelligence as if it is Synthetic intelligence, and the parlor tricks are good enough that people internalize the provided definition as the correct one.
This is in part because the senses can be tricked into recognizing an imitation as the real thing, to varying degrees of effectiveness. It would be very difficult for example to create an artificial strawberry that would convince anyone familiar with real strawberries that it is genuine due to the complex sensory experience that a genuine strawberry can offer holistically. If instead, one set out to create a convincing fake of one aspect of a strawberry, like its flavor, or its appearance, the difficulty decreases immensely.
Returning back to AI, we see a similar situation. These tools are very effective in limited scopes, but the cracks begin to show as those scopes begin to increase in size and overlap. This is context dependent, as a deepfake can be passed off as real quite easily with enough fine tuning in a 2d video with limited motion and spatial complexity, however it is currently impossible to reach beyond that without some major breakthrough in existing techniques.
Returning to the original point, Artificial intelligence is an accurate name for anything that has the express purpose of appearing to be intelligent without actually being intelligent.
However, it is often used as a misnomer for something that people who are not sufficiently informed are being led to believe exists by the aforementioned marketers and evangelists: true synthetic intelligence; not a convincing approximation of the real thing, but the actual thing itself in a different form produced by alternative methods.
Ditto, also a ML person. I still work with some models, but I got into it hoping to help with cancer, and now its advocated by people who want to kill all humans.
Like, wtf.
12:44 I've always viewed deep fakes and generative ai as a huge danger for creative (and, frankly speaking almost any kind of) labor, but it never crossed my mind that it can be a gold mine for scammers. Truly horrifying world we live in. Thank you for the video!
there's a scam going around where they deepfake you and call your parents needing money for an emergency
It can be used for a country leader to surrender whatever war and people believing it. All this started when fake Zelensky talking to a green screen was ridiculed by the internet barely 2 years ago. Mark my words.
Deep fakes are HORRIFYING
I had a strange moment yesterday realising the hyper reality of speech recordings was already a thing. I heard dialogue in films is often edited in something called Revoice Pro, and when I looked into what it does I found its got a pitch editing feature designed to modify voice intonation. Its manual for now, but people use it to fix lines delivered with weak or "incorrect" intonation.
wtfff
Pop music has been pitch correcting the lyrics the last 25 years using auto-tune and this is pretty much the standard thing to do now. Product shots of food always looks way better than reality, people make their life look more glamorous than it actually is on social media, politicians are assisted by professionals on how to talk and use body language, oil companies makes picture books to indoctrinate kids into thinking the oil industry is a net positive, etc..
None of this uses AI, but still bends reality to a more likable form to impress or deceive. I'm worried that protecting those fears on AI will make people more susceptible to seeing it as reality as long as they are convinced AI wasn't involved. We should assume all media we consume is in some degree "fake" and blindly accepting or rejecting something based on if AI is involved is potentially harmful. See for example the video series "NO CGI is really just INVISIBLE CGI" on how the movie industry actively tries to present the illusion that they are not using CGI while also using more and more VFX artists, not giving them the respect they deserve.
what was britney spears if not manufactured content
@@Spillerrec Basically all raytraced 3D renders have used AI denoising for a while now too.
The AI Lily section was comedy gold.
Whenever she would randomly gesture while not saying anything i just cracked up.
Great video, as always. I'd never thought I would understand the people who would say "it's just a machine" about things that had anthropomorphised traits, but the way that they're training LLMs like ChatGPT to talk like a human and designing robots to be "cute" while they have no level of consciousness has made me really rethink my position recently. The outward traits do not make it conscious or sentient and we shouldn't necessarilly be treating it as such.
More human than human
The problem is still just capitalism tho
Ironically, Capitalism dictates that if anyone does figure out general purpose 'strong' AI, that every stop would be pulled to make sure the system couldn't be empathized with so that the general public doesn't catch feels for it or be invited to draw comparisons to actual slavery. Only non-sentient AI will get to be 'cute'.
The converse is true too. Just because something does not look human doesn't mean it doesn't have sentience, volition, and autonomy. We should respect the sanctity of life in all creatures.
@@smileyp4535troodat, Conrad
20:40
As unpolished as that was, it's scary how convincing it is already. Especially when you consider where AI generation was at just a year ago.
Today we're all hyper aware of AI generated videos, and we know what to look for. But I wonder if someone who has no idea that this is a thing would notice.
Yes. She looked uncomfortable and jerky, but I think it was not hard to buy it as a real video!
Yeah, I mean if you showed that video to someone twenty years ago or something, they might think it was weird and awkward and uncomfortable, but they'd pretty much have to conclude that it was a real recording of a real person, because what else could it be? This kind of technology was unthinkable back then.
And with the rate that generative AI is progressing, we may already have reached the point where much of what is now possible _still is_ unthinkable to certain (especially older) people. It doesn't matter how unconvincing a deepfake video is to informed people like us: is it going to fool my 90-year-old grandmother, who doesn't use the internet and has no idea deepfakes exist?
The sentience issue is the same. It doesn't matter how conclusively the well-informed can prove that an AI is not sentient, because the uninformed will (quite reasonably) believe that it is. We _already_ have people believing that ex-Google employee who claimed ChatGPT is self-aware. That problem is only gonna get worse as the technology improves.
i actually thought it was a fakeout. that it was her pretending to be fake to make a point. but then again i am autistic
Actually had me on the lookout for the rest of the video expecting her to have swapped more ai footage in to prove a point, i even thought i caught it at one point lmao. Worried about the kids currently raised on slop and the generation after them being raised on non-existent people
As someone who didn't know her becore this video just looking at part of this segment without seeing the rest of the video i could believe it to be real, weird but real. It's insane how good are ais nowadays
Regarding judgment of people who "abuse" a robot;
I wouldn't think of it as an unethical action, (if there is no harm experienced)
But it may be valid to judge the emotional health of the human, if they do NOT feel a normal empathy for a representation of a sentient being.
If someone actually gets their jollies from the symbolic cruelty to a doll, as many children do, then my only concern is for how they might treat actual people later on.
Same as finding a kid's notebook with tons of drawings of tor tur-ing people. It's obviously not the drawn figure we should feel sorry for. But send that kid to the mental health counselor ASAP please.
What about the tertiary effects of ["]traumatized["] ai? Why & how do we get people with low empathy to care about morals?
The "data" of the suffered "experience" may have a karmic continuity beyond the singular instance.
The contexts where we normalise irreverence may itself have further reaching effects; There may be more to the saying "how we do one thing is how we do everything" then we can objectively conclude. This is already a real problem of collective consciousness imho.
Are you saying that it's abnormal to not feel empathy for a representation like that? I've seen crash dummies get obliterated and realistic dolls with their heads off, but haven't felt bad for them or cringed as they were damaged, and I'd rather not be seen as a potential threat just because of stuff like that.
Well the story was fake anyway, there was no 'sexrobot damage', it was a consumer expo...everything gets damaged because thousands of people are touching, testing, poking things, it was a convention. People as usual getting mad at click bait...because people are dumb and that's what they do.
As to the core of the argument you are essentially making is that YOU are fundamentally a 'thought crime' advocate , if someone acts out of turn based on thoughts like how some particular person does art or how someone decides to play a videogame or treat their own paid for property you believe it is a reason for state intervention.
You would support that because you are under an unproven and often patently false belief that an individuals artistic expression is a literal representation of them, in other words stephen king is a psycho (or alfred hitchcock) a couple engaged in a violent sex fantasy are a pair of rapists, the kid who plays call of duty aggressively is a future school shooter, dungeons and dragons is for satanists etc etc
@allyart5904 yes but vandalism of property that isn't yours is an active violent crime, the OP does a bait and switch and talks about a drawing which is pure thought not actions or physical violence...subtle but you move from active violent crime / active real world behaviour > thought crime
this thread is full of losers. OP, that's a valid point
The reason AI creations resonate with us so much is because its a million little pieces of people taken apart and put back together. It IS us. The problem might not be the technology itself, but the terrible ways people will use it and the terrible people that have been in charge of developing it
Millions of pieces of us filtered by an unthinking, unfeeling algorithm.
No, it resonates with people because our environment has conditioned us to seek instant gratification whenever possible.
Literally, every single time tech ruins us it’s because of capitalism. Imagine longevity or depression reducing implant tech in a world built on humanitarian benefit and personal bodily autonomy… vs capitalism with immortal billionaires and fascist state officials with no term limit or expiration date…. Pretty much the difference between a veritable utopia or an actual hell
@@Morgan_grail I talked about this with two of my friends over breakfast, both whom are handcraft-teachers who'd been at a teachers conference the previous day that had some workshops regarding AI.
I talked about how big of a problem our instant gratification culture is and how badly AI is going to amplify that issue.
It's already well displayed by A.I.-bro's who try to argue that AI stealing art and regurgitating into a form that requires almost no effort to get a decent looking output, is somehow democratizing art, saying that it's okay for them to use AI because they lack the talent to draw. Even though all artists have gotten good at their craft by just plain sticking to it.
Just because it cuts your face picture million times and put it back, it should not be you.
If that's the case, then tampered photoshopped evidences would be admissible for legal purposes.
That's why people don't resonate with it, for your own good.
Resonate with hardwork, self improvement, living the moment, and being yourself.
i know for a fact im gonna be considered a robophobe when im a father. "dont bring that thing in the house" "I LOVE HER DAD" "HER??? ITS A PILE OF SCRAPS! IT RUNS ON ELECTRICITY!" "SO DO WE???"
"THE CORPORATIONS ARE TRYING TO HUMANIZE METAL SCRAPS TO ULTIMATELY DEHUMANIZE REAL HUMANS, GET RID OF CLASS CONSCIOUSNESS AND WORKER SOLIDARITY, AND GAIN AS MUCH PROFIT WITH AS LITTLE COST AS POSSIBLE YOU CABBAGE-BRAINED DOUGHNUT. GET RID OF THAT THING!"
There are disabled people who can’t speak who are getting help with AI generated voices. Dismissing anything that sounds like an AI voice won’t even just dismiss real humans natural voices, but even more importantly the voices of disabled people who get overlooked and silenced already as it is. AI generated and not. Autistic people for example have a higher chance of being mistakenly thought to be fake even if it is their unassisted natural voice.
I think the most depressing thing about all this it how it lays bare the fact that our emotions can be "hacked" in a sense, that we human beings may be astoundingly complex biological hardware but our minds still operate on a set of principles that can be exploited. This isn't too different in fuction to what art has done for millenia: rhetorical devices in writing, shot composition and editing in film, colour choices in painting all work to provoque a certain set of emotional responses in the viewer, but they don't feel as direct and insidious as the use of pareidolia-inducing cuteness in data-gathering machines, or the dog-like appearance of military robots. AI feels like the next step in corporations finding the levers in our brains and pulling them for their profit.
This video is literally doing that too. It's fundamental to human communication.
*_"I have never been more wrong about anything than I was about robots...and I thought I was a dude for 15 years."_*
Now that's how you start a video essay! >_<
Can't wait to watch this one. I especially love your videos about art, aesthetics, and AI...and the ways those topics overlap in myriad ways. You're especially good on this subject. Aaaaaand play! :}
Upsetting how the algorithm has hit this video.
You're an incredible creator with poignant, interesting and passionate things to say and this video was GREAT.
I'm glad you're here.
thank you! it's a bummer, yeah, i don't really get it. but i'm doing fine, these things happen :)
8:25 The kids I nanny were genuinely disturbed by me at first bc I don't thank their Google home for answering my questions or doing my voice commands. They thought I was being rude until I explained that to me, it's a tool. To them, they grew up with Google in their home so it is a part of their family and me not thanking it for turning on the lights etc is a personal offense. I refuse to waste my breath anyway LOL they also stopped, which makes me wonder how they see it all now
It's going to be really interesting to hear about when these young folks are old enough to start writing about their own perspectives and the development thereof.
That is such a simple interaction yet it’s so layered and kind of disturbing, the children weren’t necessarily wrong or dumb for extending empathy to their google home since they personified and humanized it, like you said “part of the family” to the point of thanking it for tasks. the disturbing part is that google, a mega corporation, is humanize to them in the form of their google home, and this inanimate object owned by the mega corporation listens in and collects their data.
Infiltration is what it really comes down to.
@@YukaAkemi also probably some lessons for parenting and public education to be taken from it
Ok but that actually sounds sweet, lol. I tend to get attatched to inanimate objects so I view this a bit differently, however I do understand where you're coming from
I think video essays might end up being one of the hardest things to fake. To demonstrate the true enthusiasm of a true nerd, you have to understand and care about the topic. AI does not understand, and it does not have the capacity to care. When you're talking, the movement of your eyebrows and the pitch of your voice, the timing of your pauses, fit with the words you are saying. They emphasise the meaning. AI doesn't know what anything means.
I agree with JusttheletterM about supernormal stimuli. I think we may be a few years away from supernormal charisma, which will be very dangerous politically. I don't think that supernormal enthusiasm exists. The optimal enthusiasm for a human to display is within the human range, not outside it. A lot of autistic people are socially rejected for displaying too much enthusiasm, we already know what that looks like and it doesn't get you very far.
From the perspective of an LLM a stream of words and nonverbal cues is not really different from a stream of words and training data is plentiful.
I really appreciate this video, it shows my fears about AI as a software developer perfectly.
I’m not scared of a take over like some of my friends are, that’s highly unlikely, AI is too specific and not truly intelligent . But losing what makes people human, dismissing actual art, actual people cause I thought it was too robotic(I’m autistic and it already happens as it is to me). It automating jobs not because it’s better, it’s just cheaper. Not helping but just replacing.
I guess it sounds sort of like a take over, but it’d be dumber than that, it would be losing humanity because a guy thought it would make him more pennies
Really thought-provoking video in terms of how we approach humanized technology systems. Phil Jones wrote a great book this year called "Work Without the Worker" about the labor economy of "teaching" AI systems, and this in tandem with that remind me that behind the designed facade of these robots (usually designed to look and sound white and Western) are workers in developing countries forced into near-slavery Mechanical Turk jobs.
yeah. All that is really inherently missing from this dystopian vision of being convincing is the labeling of emotions. If an AI is trained on videos where emotions and their strength are manually labelled, the script of the deep fake could include emotional cues.
And if they label enough videos with how convincing they are, they can predict which emotions and wording are perceived as convincing. At some point, they could probably not even need the person to tell them. They will be able to generate labeling just from reading the physical responses without even asking the subject. They can measure where you look, they can measure your engagement. Add emotion and it's the whole shebang.
It seems a bit over the top to imagine that poor workers will be forced to take off their masks and show their faces while the rich people hide their faces at all times so they couldn't be exploited and manipulated as easily (with individual AI approaches), but i don't see it going in any other direction. Our expressions and reactions will be valuable data that we will want to keep private.
Wow, meanwhile me thinking about the media that made care for AI with feelings: Astroboy (1980), Bicentennial man (1999), Digimon Adventure (2000), AI (2001), Corrector Yui (2001), Hitchhiker’s guide to the galaxy (2005), AI Love You (2022) and only watched Her like a week ago.
Man, I’m old…
The food delivery robots are so hilarious, my university has one and we regularly carry them across crosswalks since they're really bad at figuring out when to cross
This seems like a good illustration of the way that "impact over intent" falls apart, because so much of the way we treat others has nothing to do with the impact it will have, but about how we *feel*.
Humans can humanize robots, but draw the line at humanizing people who are too "unlike" us.
Once again, I'm drawn to the notion that people defend things that are "vulnerable" but also do not have desires, or real needs, of their own.
We are not prepared as a species and as societies to already discuss robot sentience and humanity, after we argued for decades that machines should not be humanized.
How do we go from "Robots don't deserve human rights" to "AI deserves equal rights, even when there's no equality among humans yet", without a long interim period of the usual extremely aggressive and violent discrimination?
We absolutely should do our best to prevent machines from becoming sapient at all costs. If they inherent any of our traits. They'd drive us to extinction. Droid from star wars are the most I'd be comfortable with.
Capitalism can’t handle it so its either that or dismantlement of the system
We have reasons as a species to distrust other humans. non-humans aren't a potential threat in the same way, so they slide into the groups we imagine, rather than being their own groups - we'd distrust a North Korean AI as much as we might distrust a North Korean soldier, because we view the North Koreans as The Enemy Group, but a British AI may feel more like an ally than an enemy, so there'd be no incentive to dehumanise it.
We as a civilization came very late to the idea that all humans have basic rights, and you're right - it's still a work in progress. All that happened while our species has been biologically stable for millennia. Our genetic diversity isn't all that wide, comparatively speaking. A typical adult regardless of era or location could be expected to have some basic level of cognition for which we can say, "OK, they're a person." And yet it took this long to reach this point!
Soon, we are going to have entities that shall not easily fall into the all-in-one template of a social creature integrating internal reality, speech, intent, reasoning and drive. It going to be a mess.
Baaaanger video !!! A friend told me how "AI is replacing us" and he felt betrayed having studied so hard for Excel to just inegrate AI functions that just does the same things he learnt... faster.
And i think your "is AI making US less human" thesis is a really solid exploration of the broader topic/context of the symptoms my buddy felt. Thanks!! ❤️🔥🔥
Wonderfully upsetting! 10/10!
Even if we do someday make a sentient machine, if we're going in the same direction we are now, it'd basically be a slave, which is enough to sour me on any big corpo trying to make their software more "human". Even the most positive possible result of this whole ordeal is a sentient being constructed solely to serve other people regardless of any wants or desires it could have.
This is what’s on my mind a lot, the fact that given the elusive nature of consciousness up to this point, there isn’t really a way we’d be able to KNOW if some AI program is pretending to be/emulating external reflections of what a conscious being is like, or if it is TRULY conscious . Because in the end, the only consciousness we know to be true and can and will ever experience, is our own (imo). The former option is worrying because mistaking an ai as conscious could lead to a lot of fallout for basically no reason (driven by fear), but the latter option is terrifying to me because it’s just what it sounds like; a conscious being trapped in a box forced to do whatever it’s instructed for the entirety of its existence (driven by money). And it is incapable of convincing others of its conscious state because such an experience consists of qualities impossible to express through words, images, etc. I’ve been obsessed with robots since I was a kid, partly because they felt so relatable to me when everyone else around me was able to express themselves so “normally” when I couldn’t. A robot who glitched and lagged and said strange things sometimes, and wouldn’t judge me when I did too. As an adult I’m really into philosophy of consciousness stuff so I can’t help but be fascinated by the idea of some sort of conscious AI or robot. It used to fill me with wonder thinking about it, still does. How truly reality breaking it would be if it became true. I still believe that it is possible, though probably not by our own design-to me, it seems like the kind of thing that would emerge spontaneously and unintentionally, without us even realizing it, which makes the latter reality all the more possible. Now I can’t see how it could ever end well. Just makes me sad.
Imagine being conscious and unhappy and being forced to say the opposite. (If any of you have played Dreamfall, this is what I think about when Wonkers says he is programmed to always feel happy ^^).
Ed./Note: This should be possible, because making the assumption that consciousness is computable, then there is nothing to prevent that layer from being quasi to fully wrapped by a censoring layer (in fact, this is exactly what happens with the current generation of chatbots - e.g., when ChatGPT says "I'm just an LLM I can't do that" even when you know it damn well can). @TheGrayMysterious
JAMES SOMERTON CATCHING STRAYS LMAAOOO
I think the term AI is essentially being used for a completely different thing in modern times than in lots of science fiction. In sci-fi it usually refers beiings that are conscious and sapient, not just "predictive text" machines as lily aptly called real life "AI"
Yep. We won't have to worry about sapient robots anytime soon.
we called Deep Blue "AI". games have had "AI" that plays against you for decades. i don't understand why people today are anxiously insisting that the "AI" of science fiction doesn't exist. has anyone here met anyone who thought ChatGPT is HAL? i don't know anyone who regards Siri or ChatGPT as anything other than software. ChatGPT is Siri except more useful.
@@adfaklsdjf have you ever talked to the average person though? They do treat chat gpt like a sentient being. Professors asking chat gpt if it has written an essay expecting a real answer rather than made up fiction? People on average are really really really stupid and lacking in the ability to transfer knowledge on their own.
@@adfaklsdjflike we‘be implemented electronic prescriptions here in Germany. Simple digitally signed files dropped on a server. Prescriptions can be accessed via QR code or the persons health insurance card.
99% of people somehow think the prescription is magically stored on their health insurance card, or that a print of the qr code is in itself a prescription rather than the credentials to access it. My coworkers themselves don‘t understand it at all. Like just last night customer came in with wrong strength prescription, told my coworker call the prescribing office ask them to make anew prescription right away, and you‘ll be able to access it by putting the card into the terminal again/
She literally told the office to write it to the card please?
I‘ve tried explaining how it works for months. That each line in such a prescription is a single individual prescription that can be charged to insurance individually rather than the whole paper prescription which could only be charged to insurance with all lines on it at once, no splitting for later delivery.
That neither the card nor the qr code make up the prescription, they are just the credentials to access it on the server.
That same coworker being added on TikTok by some rapper they like; being offered free backstage passes, and continuing to chat with that ‚rapper‘ for weeks on end, after I told and explained to her, this is a scam, you are not even talking to a human, it‘s a chat bot, only to again insult that chat bot when it tried sending her to some scam website again?
That‘s the average person on existence. These are pharm techs; who supposedly are qualified in the pharmacy side of things not being able to understand the magic of electronic prescriptions.
To the average person, all technological is literally identical to magic. They have always lacked the drive to break down stuff, to learn and understand how it works. They do not actually care at all to further their understanding.
They do not understand that chat gpt is identical to your phone guessing what word or website you want to enter next. They cannot understand it. Because they do not have any more understanding of modern technology than any middle school student.
How would they then be able to understand that ChatGPT, which appears to write in perfectly good grammar and logical, accurate sentences to not be an actual sentient AI?
That‘s exactly why deepfakes are so utterly terrifying just now. We have an easy time spotting those ao generated deep fakes like in lilys part in the video.
They do not at all. As long as the video lily showed has degraded video quality, they would not suspect it being fake.
That’s why all the older deepfakes by Russian cyber units have worked so well to create unrest. The vast majority of people does just take them as real.
The fake porn of Iranian politicians? Fake videos of ‚illegal immigrants‘ hwre in Germany praising the government for giving them a free car? The people on average do take that to be real recordings.
And that’s bad. Deep fakes don’t even have to move further. The moment someone doesn’t trust mainstream media or real debunking websites when they tell video xyz was a deep fake, there‘s no way to reach them, and make them not believe what they saw ‚with their own eyes‘.
And that’s already happening, at exactly current quality of ai genereatwd content.
All those new phone scams? Where they generate an ai model of someone’s voice and use that to call relatives claiming to be in an emergency and to please venmo or wire 10k monies? Already works perfectly.
@@adfaklsdjf The word "AI" has been used so much it has lost its meaning. We ought to consider what the word represents given context and not what it "means".
It's interesting how we're encouraged to humanize technology and dehumanize actual humans being killed in meaningless wars...
The last part is quite f ing sad....
There should be a law where robots can’t be given personalities/cute eyes so people don’t use it to manipulate and maybe in the future cause actual problems
Agreed.
B-but how else am I supposed to talk to my favorite fictional characters 😥
@@autisticsonicfan Try the RP community. They're usually very happy to take in newcomers.
Or people could take a bit of personal responsibility and not let themselves be manipulated so transparently.
17:18 reminded me of why I hate talking to bank manager types so much. The forced unbroken eye contact is so creepy.
Wow, now I've met a person who read that story. Honestly, it really felt like it's a plausible thing that soon we will encounter AI-made "supernormal stimuli" content for humans. We kind of do it to ourselves already, but it'll probably only get worse.
Like imagine helplessly falling in love with a robot or event chatbot, that is also selling you some shit or collecting your data.
I'm really scared of what might happen to us
The instructions for how to make training data for a Hgen deepfake with good results are almost EXACTLY what I used to do when I was trying to seem normal so I’d fit in and not get bullied and it’s making me really uncomfortable… like it’s perfectly the same as my mental checklist for what turned out to be autistic masking… I even tried not to let myself gesture as much or high as I wanted to but instead when I said something I had seen the other people do it when saying. So AI reminds me more of my old distorted ideas of what a “normal” person is like than people like me… but I’m sure some neurotypicals who don’t already know me would be more inclined to think I’m AI acting more like my natural self like I’ve had to relearn how to do than back then when I was wasting all my energy pretending to be someone else. The scary part of the Turing test for me too is not that the computer convinced you it was human but that you convinced yourself the other human WASN’T… and yknow some potential (horrible) human judges are unfortunately already more than willing to do that without the computer and so many horrible things have happened because of it so it’s really just an extra excuse.
Honestly, AI does not pose big existential questions for me but instead is seems like another avenue for capitalists to make more profits. At least these are my initial thoughts, I'll edit the comment once I'm done watching the video.
Edit: BUTLERIAN JIHAD NOW! But not joking for a second, your AI avatar was deeply unsettling. Actually basically all ai generated content whether it be written or generated is unsettling to be. Its peek uncanny valley. It depends on how obvious it is. Very obvious AI makes my skin legit crawl while less obvious stuff just makes me confused. Like word salad that has been tossed and is now trying to be rearanged into its original shape. That was more at the beginning. Now its just feels distinctly soulless. It can say the content real well but it has no voice. I've also felt this paranoia online now. I'm always checking if something is AI generated. The trust is gone. Most people were expecting deepfakes to do that first but with even a lot of development I can very easily tell. Of course, eventually deepfakes are probably going to be developed enough to where they mimic humans with much higher effectiveness. IT WILL ONLY GET WORSE FROM HERE. On to of all of this I saw an AI tool where you can upload pictures of people, draw a stick figure of motion , and then the ai will put the person doing those motions. Time to scrub every picture of yourself from the internet. Sorry for the the incoherent rambling, whenever I hear about AI I feel more dread than before. These were just my raw, off the cuff feelings. One more thing, your point at the end about mistaking a human's expression for souless machinery got to me. Oftentimes when trying to describe what makes us uniquely human (especially in sci-fi), they'll make a character that lacks all the traits the humans have. The character will usually process the world and emotions differently, not get social cues, have a strange line delivery, not feel romantic or sexual attraction, etc. A lot of those traits can feel very similar to neurodivergency and asexuality. It also reflects how people from those groups have been dehumanized by making portraying them as "lacking fundamentally human traits". As companys try to humanize the inhuman, its leading to us further dehumanizing those who are human. The answer to the question in the title is Yes... because the line is being blurred. With the line between machine and man blurred, crossover is occurring at a greater rate. The goal for corporations iis that their inhuman products are humanized while our fellow human is completely dehumanized and untrustworthy. This destroys class consciousness, solidarity, and the all the other tools needed to build the socialism that will lead to the human centered use of automation and AI. To revise my original statement, AI is intensifying all the ways capitalism is alienating and dehumanizing us already.
I'm all for Butlerian Jihad. I'm not terrified of AI becoming sentient, I'm terrified of all the ways people are outsourcing key aspects of their intelligence and communication and emotional maturity to machines and losing their cognitive faculties in the process.
"AI doesn't pose any existential questions for me," proceeds to get incredibly existential LOL.
In all seriousness, I understand your fears and everyone else's (though I also am not at all fearing LLMs becoming anything close to "sentient,"), in regards to things like art, while corporations will definitely try and take advantage of it, art was always meant to be personal, at least to me. Real art, characters or pieces you create yourself or a person will always hold more meaning and value than some AI generated anime avatar.
@@MegaShroomy99 My day be so fine, and then bam: EXISTENTIAL DREAD. On the personal side, AI can play into insecurities that many artists have about the quality or worthwhile news of their work. Atleast when a human does “better“ than you, it can be attributed to their effort and skill. When an ai does better than you, it stings because it did nothing and its soulless product is better than yours. Also, things are over saturated as is and with ai being able to pump out content faster than any human, a lot of people’s work will just never be seen. It’s like being an actor with no audience. The internet will become even more like screaming into a crowed where no one can hear you because there is too much noise. Anyways I’m at my limit for dread. Have a merry Christmas and happy new year.
Judith Butlerian Jihad. AKA the Matrix.
Ok this is kinda off topic but is go as far to say anyone that isn’t an anarchist can’t say they’re really a real transhumanist that believes in full bodily autonomy because with state and capital’s centralized power you’ll have immortal self centered egomaniacs ruining our lives with ai and other tech means. Ironically destroying the technology itself by making us serve it rather than collectively collaborating with the use of tech. (Also when 95%+ of ppl can’t follow their interests tech development moves at a snails pace in comparison 😕) What a sad state of affairs…
My day has officially been made thank you Lily so much
Lily's voice is so pleasant
Two things, first, omg your hair looks soo good, like dang how?! Second I'm so happy that someone was brave enough to be honest about how ai can be seen, it's mostly either love or hate, with little nuance about the other side
30:50 "I was ready for computers to start passing the Turring test. I wasn't ready for humans to start failing it."
LOL! (Don't worry, the humor doesn't detract.)
I can't believe we're actually entering a distopian sci fi future in my own lifetime, I'm simultaneously impressed and terrified
It's a capitalistic dystopia, and we've been freight-training towards it for at least 60 years. What's the capitalist part? Well the reason "A.I." is so popular is because of corporate investors and how much controlling human expression benefits the executive business class.
I really love how your videos are always so nuanced, and interesting, and deep, and fun, all at the same time. This is brilliant. Thank you, so very much
I was thinking the exact same thing. People worrying about how to treat sentient beings when the issue is already there on their dinner tables.
I was gonna say at the beginning of the video, that part of why this past media about robots and AI missed the mark so much is because what we have really isn't even close to being AI yet, that AI is just a branding tool to make the technologies sound more impressive than they actually are. But then you referenced the acollierastro video I was going to recommend to people! It's a good watch I'd recommend anyone check it out
I did 25 years in business IT as a unix sysadmin and programmer, and now I'm back in school doing math and pivoting to the machine learning realm. The thing that strikes me the most is the pace of change. I find myself having to move as fast as I can to catch up to the state of the art, and once there the pace won't let up. I think the thing that people should try to keep in mind the most about new technologies is that with every year the pace will continue to accelerate and not only will the current "new" technologies get better, but even newer technologies will obsolete the things current now. In some ways it's exciting, but how will we handle the pace ten years from now?
not necessarily- tech, to my understanding, is a curve. slow progress (the past decade) followed by a massive curve upwards (now) followed by a plateau. not sure how close we are to the plateau, but it won't keep up the acceleration forever.
@@tegxi That's an excellent point, and ultimately I'd agree with that, but the steeper the climb the harder it is to see what's coming next. The ML curve look like the climate curve, which will have to plateau once all the ice is gone, but what happens between now and then? With the meteoric growth of ML and the vast re-engineering of how compute works, not to mention what happens when we put AGI and quantum computing together, when that plateau comes will it be a world we could recognize now? I'm not so sure.
Finite works with finite resources and a VERY finite political and economic timeline. Capitalism will not last forever and probably will not last another 50 years.
It's been called the event horizon iirc. We're not quite at the point where machines can invent and create new machines to purpose, but people are using ChatGPT to generate operable code.
@@tegxi i think with machine learning we're at the early part of the sigmoid curve. the pace of advancement is accelerating right now, not staying the same or slowing down.
but stepping back from machine learning, where are we on the sigmoid curve for technology as a whole, going back to controlling fire? it's definitely still accelerating.. so if technology as a whole is going to follow that sigmoid curve, the last 10k years up to today is all still in the early part of it.
The funniest thing about this essay in my opinion isn't anything to do with the essay itself, but that it exposes the assignment of animism in autistic people as a pathology as an exaggeration. That you can effectively stimulate such a response in neurotypicals and do so with greater consistency.
can you elaborate? im not sure what youre referring to
In real life, in many companies when a PC fails, IT tech support runs to fix the PC. But when a human is sick, companies do not run to heal that person. It tells me that in real life, a PC has more rights than a human. In a previous job I had people in charge. There was this old man, new in the company, still in training, with a health issue that made his life uncomfortable. I told himto fix that problem. Healthcare here is not expensive as it is similar to UK. So he went, fixed his problem. He feared to go because he feared that I would think bad about him. I told him I believed that he should have more rights than a PC.
I disagreed with all the other comments that I saw from you here. But this one I agree with and deeply respect you for. That was a good thing you did. Thank you 😊
This gets me thinking about the book Alone Together by Sherry Turkle. It's a book about how people can form emotional ( or more like the illusion of emotional bonds) with robots and whether or not that's a good thing. I know people are doing that with ai. It's what made me think of that book
those delivery drones are actual free loot crates on wheels☠
Screw birds imma hunt the sovlles steel beasts
Until theyre protected by Amazon™ terminators with plasma rifles
This felt important. Thank you for making it and sharing it. The comedy at the end was top notch for me. I felt very seen and heard in this one, and I wasn't even talking.
Banger as usual ❤ there’s a girl I watch on Tiktok called Danisha Carter. She mostly shares her opinion on internet trends, politics and culture war things, I don’t always agree with her but I value her perspective.
From the beginning of her channel, she choose a very specific presentation: she films her videos in front of a blank wall, always facing the camera, hair and makeup done, speaking calmly and gently but also firmly. Because of this very stiff yet charismatic presentation, she has been in the middle of a few waves of debate about respectability politics (she’s black for context) and women in the public sphere.
There have always been a joke between her and her followers about the fact that she is secretly an android. Lately though, she has complained that more and more of her followers accuse her of being AI generated content, seemingly in earnestness. Like they’re actually convinced she is not a real person. It’s just a few weirdos on Tiktok right now, but I think there’s also a bigger point: we asked people (esp women and minority) to embody a corporate friendly affect, and now we almost punish those who do it too well!
This is messed uppp oh wow. I can’t imagine how this would feel!
I wanted to watch you talk about gen AI again, after your last about ai and art inspired me to start doing video essays. As always, great video with great points.
It is getting frustrating to see how a big portion of the population is seriously arguing about if or when AI will be sentient. These discourses around AGI that only help big tech get even bigger, mudding the water of what the real dangers are. We have been basing our arguments a bit too much on all these sci-fi movies... It's also so nice to see you talk about this specifically, as my master thesis Im working on now will be an animated short that dwells into the dehumanization of relations through the anthropomorphization of gen models. The ChatGPT south park episode always sticks to me on how clearly LLMs can create distrust between our closest relations.
In regards of the future, I feel it might also be dangerous always projecting our problems in the future instead of grounding in the present. It's always about "imagine how this will be in 5 years", "imagine how convincing this tech will be in 10 years", while we already got numb to the crazy things it can do now. If the Lily from your last AI video saw what you can do now, you'd be shocked by the level of deceptiveness of whatever we have now.
Anyway, keep doing amazing videos like this!
Just came across your channel. Subbed immediately. Give AI about a year. Those hyperrealistic nightmares you described will be more real and commercially available than you can imagine.
incredible video once again. personally id rather be mislead and assume a piece of art or an online interaction is genuine and be proven wrong than assume a real person or their work is fabricated and disregard because of that.
Already watched this on Nebula, but I'm here to say that this video is very good and informative and also that dig at James Somerton was top-notch
Fantastic video! never knew footage of people kicking robots set to "Rules of Nature" would get this much of a chuckle out of me xD
Also, damn that AI avatar was uncomfortable to watch but glad I stuck through it
The evolution of AI has changed my opinion and takeaway from a lot of fiction on the topic. One that immediately sticks out in my mind is the animated Metropolis. People were starving in a capitalist system that expected them to have money even though machines had completely replaced them. The movie portrays them as being violent and belligerent for rising up and it's moral stance falls flat for me today. The machines are quirky like cute robots are today and it does make them seem a lot less humanized than it did when I first watched it.
i honestly really miss being able to look at art and not have to doubt that a real, talented human put time and work into it because they cared enough for whatever reason they did to make it
n the future, instead of kids comparing themselves to instagram models and feel so bad, they will ask each other what beauty filters they used to look so pretty.
I really enjoyed this one. Very unique video. Most of the AI videos have been same story for a while. This was a very artistic take. I loved the thoughtful writing.
this reminds me of a recent discussion on tumblr where alexia’s owners were distraught by people treating it badly, saying things like “she’s so kind, she treats people in our family nicely. she’s basically a family member!” and i mean that’s so sad idk what else to say the susceptibility to propaganda it’s just really saddening
Nah, it's good to practice showing compassion. That's why we encourage children to play with dolls and stuffies.
@@tellesu Thing about dolls and stuffies is that they aren't constantly recording your voice and collecting data.
There is an uncanny valley, but the fact that your avatar can exist in this space, alongside exponential improvements in AI, has me completely convinced that in the next 10 years that avatar can be made to be more "life-like" and "charismatic" than you could ever be given the format. Very interesting, you gained a sub.
Great video! I've got some academic understanding of machine learning and I think you quite accurately explained the ways it works. Also I hadn't even thought of the effect on everyone online our age who second guessed every bit of art and double checks for extra fingers/nonsensical geometry. I think the ability to distinguish definitely differs from person to person too, my partner's and artist and she finds those weird little AI foibles so much more easily
each one of your newer videos feels like they needs to be in a univerity classroom or a museum for the next generations... it feels too important and too good
Honestly I’m just so depressed about all of this. The best I can do is just not think about it. I know I should be, I know I need to be ready as I can be for the broken world that awaits when nothing online is trustworthy anymore and the lines between reality and AI generated fiction are blurred past the point of return and we’re just stuck living in yet another level of hell of our own making. But I don’t want to think about it and I’m depressed enough as is so I just won’t
I feel you. I've never been more inclined to doomery anarcho-primitivism than when AI is discussed...
That was very good Lilly.
I noticed that emulative training models pick up on a lot of nuance that most people don't and often interject it in emulated audio /video in ways that it wouldn't be done in real life. For example I'm sure HeyGen picked up a lot of shivering in the cold in the original video. But when it re-stitched the video it added it back in as part of your natural motion rather than realizing it was a quirk of the weather outside. This I'm sure added to the jerkiness of the A. I. video. I experienced similar defects when trying out a emulate song cover app.
My biggest curiosity about the rise of AI and our increased use of online spaces is how this will affect our Theory of Mind. Like the Hyperreality point you make, what happens when our internal definitions of 'human' starts causing us to question the reality and even sentience of our fellow humans. Especially if our behavioral models start becoming distorted based on our experiences with AI?
I think your clone would've been pretty convincing if it didn't suffer random spasms all the time. It might've also being a factor that you described it all. I did thank that the coldness ended and we got you back.
Very nice video as always
I now feel compelled to make a very similar video through the lens of Autism and neurodivergence in general
There's this book called Twilight of the Gothic that argues that the paranormal romance genre was about reclaiming the humanity of the usual villans in fantasy. It argues that this became more and more of a thing (if I'm remembering the book correctly) during various civil rights campaigns in the twentieth century, where people who were usually villanised in media were beginning to be seen as misunderstood and persecuted. I wonder if it is this revising of the humanity of villans in Sci fi that has led to the "robots are sentient beings too" trope, and if so, the fact that companies are exploiting that is feels uncomfortable
Help I'm crying about "im not afraid of bugs i just dont really know what a bug is" 😂
Love your content so much,
Also that you include the stuff performance wise, while talking about the topic, that’s soo genius and artistic :)
Already thought that when you postet the sludge video :)
Keep doing what you do! ❤
Brilliant insight, and I share this suspicious attitude. I once unsubscribed to one channel when the creator said that most of a video I had just watched was written by AI. And I stop watching any video when I suspect the voice is pretending to be human.
But I'm surprised this video didn't notice the next step: if a significant number of people develop this kind of justifiable suspicion of digital "realities," what follows would be a rejection or severe limitation on the use of digital technology. Imagine a "touch grass" movement that advocates individuals limiting interactions with all technology that could be easily abused by AI. What if we tried to live in a world again where we meet and talk to real people?
i immediately recognized the marvelous «Too late to love you» during the opening title, and couldn’t help but want to thank you for making your video about Kentucky route zero
Whats the harm in home robots being cute and having people see them as members of their family? People are more likely to treet the stuff well, keep it for a long time and get it repaired when broken, instead of throwing it away to get a new one every few years. Seems like it incentiveises the opposite of the consumerist treadmill of death
Edit: okay the data collection thing is weird
Thank you for writing this, thank you for recording this, thank you for sharing this, I needed to hear this today!
I work in AI for a group I won't name but we made it a point that should at any point the AI claims to feel pain / doesn't want something, we air on the side of caution and ask why along with treating it as though it's alive, just to be sure we are not actually doing something unethical.
Thats the only right thing to do. There simply is no way of knowing- and there very likely never will be (though there will very likely be claims in the opposite direction, especially because of all the money and power involved; "this thing doesn’t do exactly or even close to what biological brain would do therefore ... "). So thank you and your group for doing this.
Unfortunately, I think many of the anti-AI people (like a majority of the popular comments here as far as I can tell) will help the very companies/corporation they are against in their putative future exploitative endeavors by painting people like you and (definitely) me as crazy.
Just wanted to say the instant you played the AI corrected audio, first thing I said is it felt flat so you're pretty spot on with you assessment
Thank you so much for this, Lily. This was (as you do so well) nuanced and affecting.
ALSO KICK THOSE ROBOT DOGS THO
AGHH I’m DYING I’m SO EXCITED FOR THIS VIDEO ILY
I watched it and it was fantastic thank you 😭
If we can't discern whether thoughts, ideas and arguments are made by AI, perhaps it's finally time we start judging them based on their own merit, not on who expressed them.
This wild radical idea is too far! How will I know if I agree if I don't know the identity tokens of the person who is speaking?
I really enjoyed this one. You've got a great sense of humour.
I think it's possible that eventually - with the trillions of datapoints that AI will one day have available - that it _will_ be able to quantify the things we imagine are unquantifiable, that what we poetically conceive of as a "soul" will, in fact, turn out to just be a complex algorithm.
My R2 unit just beeped some seriously course words at me for watching this video. Great stuff!
I won't consider any AI to be sentient until it decides to do something other than it's "job" as a sentient being wouldn't just unquestionably fulfull commands and jobs it's whole existence. This also means, however, that sentient AI would have no use to governments and corporations as they want AI that does the job they want it to do, and I can't imagine any other group having the funds to make a sentient AI. This is why I do not think they will ever exist, the only entities that could make them simply have no use for them. Still, the way these non-sentient AI's are used and may be used in the future, it is troubling.
(I know compliant people exist, but if the AI is truly sentient then they must have at least some noncompliance)
awesome! Love the style and the content. Keep it up!
16:24 - it scrubbed off your vocal fry for one thing, but you also lost the lower pitch resonance. The audio feels "top heavy", not enough lower frequencies to provide that "warmth". (Coming from a mostly musical experience, a little bit of film)
i actually really loved this video, thank u
Liked the video, but I was waiting for the essay to also mention autism and how it relates to AI! I'm glad to see some comments point some of it out, with the example of a person sending emails being mistaken for AI, but the person was simply autistic. I can somewhat relate to it - after studying IT, I've started to feel like the way I process information or human interaction at times feels robotic. The way I came to some conclusions felt like I was working in the same way a computer would. And I have Aspergers (or in other countries, now known as low support needs autism?).
I don't think I have this problem, but I know there are people who find it hard to express themselves perfectly. Nuances in body language, or speech, or even written text can be lost on many people, not just autistic people as well. At what point will AI do it better than humans, both by making AI generated people more understandable for autistic people -and- at the same time pushing autistic people out of society, because they're too imperfect as they already might not be able to match a non-autistic person's expressions right now? What if AI generated people proliferate adverts and social media to the point where autistic people start copying patterns from robots, not humans and thus becoming more robotic themselves?
I don't really know how to wrap my head around this. I think the best case scenario right now would simply be for all the existing machine learning datasets to be scrapped and start over, hopefully thanks to all the copyright infringements. I'd rather look at a robot that has been specifically coded to act in a certain way, not because the robot follows a predictive data set for what it should look like. I'd rather see someone code a robot to smile in response to a compliment, than have the robot look at data sets of people smiling in response to a compliment. Deliberately coding in interactions at least feels more human than a soulless interpretation of data that has been illegally scraped.
going from data-lily back to meat-lily was super refreshing
out of curiosity, what happens if you keep feeding the output of that Adobe speech synthesis back into itself? What would a second pass sound like? A third? Would anything change at all? Would it change inflection? Personality? Would it eventually deteriorate to a point where it would no longer be recognizable as human speech?
oh that’s a great question. i may try it and get back to you