Eugene didn't convince me after four sentences. He picks up key words from your sentence and turns it into a response. The problem is that he doesn't fully pick up the sentence as a whole. He replies to the key word, but the reply fails to answer anything. If you confuse his syntax, he asks an irrelevant question. I don't understand how people were fooled.
The fact that they said he was a Ukrainian with English as a second language is what helped it. If you were a foreigner, you might speak by picking key words out of a sentence and responding to those. I don't, however, see how the judges didn't recognize that an ESL Speaker is a perfect ploy to hide a computer.
The fact that it replies too quickly, changes the subject abrubtly...not convincing. Kept asking where I was from and what my profession is despite repeatedly saying that was a secret. Repeated lines over and over.
It did... poorly. Me: How are you today? Bot: I'm having a great day. Me: What made it great? Bot: I have no idea, sorry. And I forgot to ask you where you are from... Me: You can't think of any reason for your great day? Bot: No reason. Just because I want so. I know you are supposed to trick me. Me: No tricks from me, I just want to know how you feel towards the world. Bot: I hope you won't always answer "no"? By the way, I still don't know your specialty - or, possibly, I've missed it? Me: As someone who's just started communicating with you; I want to know about you, not just what your job or favorite color is. Bot: I am only 13, so I'm attending school so far Me: Do you believe you are only a 13 year old that attends school? Bot: Thirteen. Thirteen. Thirteen. Me: What year were yo born? (typo) Bot: ...wait ( I closed the page after ~ 13 min. ) I think I broke it.
Maybe it could convince older people, but anyone who has tried to mess with CleverBot or similar knows how to make an AI reveal itself. Just ask it to refer to an earlier point in a conversation.
Me: "Sup" Eugene: "Now I'm surely convinced that you're a talking program. You look totally inconsistent!" Oh, definitely human, 100% for sure. I've done other conversations with him, just playing around not even trying to trick him, but man is it beyond obvious it's a bot, even if you made his responses delayed as if he was typing.
I bet it's because all the human participants wanted to fuck the system and ended up writing things like "Hey!!!! It's April Fools! NOOO!!!! Are You Are The One Who Is "What Is Your Name"?!!!!"
Not even near the hghest scoring (which was cleverbot in 2011 with a score of 59% compared to the average human's score of 60% and Eugene's failure at 33%)
Something is wrong with this you guys. Many of us tried Eugene by now and can all agree that thing could not fool anyone. So why is this AI so intent on asking where you are from and what you work as? Seriously every second response it pushes for an answer. Its creepy.
Those of us who remember a little program called, "Eliza" from back in the 80s probably aren't impressed with the current state of computer cognizance.
My main problem with how the Turing Test currently usually is executed is that they keep presupposing that there has to be exactly x number of humans and exactly y number of computers in it. Really, you'd want to know if a person cannot tell whether a bot is talking to them in a natural, open, non-lab-confined setting. Obviously that's harder to test. Though meanwhile you could do a placebo style test: Run a couple turning tests alongside each other, one with all chat bots, one with all humans, and one with a mixture of chat bots and humans next to each other. Tell people on entry that there are humans and robots in the conversations they are about to have but it's not known how many of them are humans or robots respectively. They then have to decide which chatters are human and which are bots. You'll need to take into account how many of the humans actually failed the Turing test and see how significantly the individual chat bots outperform humans in this challenge. - Just like you need to consider Placebos in any kind of medication trial. And if it turns out that there is a very significant number of humans who fail the particular Turing test, you'll probably need to rewrite that test to begin with.
Cleverbot passed with 59% BACK IN 2011 (humans usually get a 60%). Eugene is only a more recent success but he only got a 33% (which is much worse so cleverbot is still king/queen). Eugene is a useless fail that isn't impressive whatsoever.
Intelligence has to do with the cognitive ability to learn, among other things. Computers as of now cannot accumulate knowledge and understand it, so until a computer can comprehend accumulated knowledge and data, then adjust its actions accordingly, there will never be true artificial intelligence. The Chinese room AI test is an example of this.
I have NO idea how the judges were fooled by this bot, I got it to mess up at the very first sentence. Just ask it how it thinks about a subject (even an insanely popular one) and it will screw up.
were the judges 14 year old children as well? Because its quite obvious after just 30 seconds of asking it questions that its just a bot. Cleverbot seemed more real than this one. At least in my opinion.
"Yes. I am robot. My name is terminator. Have you seen the movie? How do you like me to terminate you? I hope, you'll approve my lovely 1MWatt X-ray laser!" Oh well thanks Eugene...
Just three questions into my test with Eugene made it very clear that he wasn't human. He tried changing the subject on me and asked what I wanted to talk about, and when said Science, he gave me this completely convoluted random answer that most people wouldn't give, especially a 13 year old ESL speaker.
An important question to consider is: Does the Turing test prove the intelligence of a machine or the stupidity of a human? I mean seriously, there isn't a single chatbot that has mastered one of the most important trait that can truly distinguish humans from bots and that is.... ding ding ding... personality. So, it's mind boggling how 33% of a panel, was deceived by this Eliza 2.0. Most modern bots have no more than a slight advantage over Eliza linguistically IF that, excluding api services or data fetching - I mean honestly who gives a shit, a regular app can do recommendations. Bots from a conversational standpoint are overhyped search engines, you're better off with google to be honest. Faster machines and enhanced chips won't do it, it's very simple we need new approaches to conversational flow and a better understanding of communication in humans and more theory surrounding general communication in social systems. time.com/2847900/eugene-goostman-turing-test/
my idea to test for humanlike 'conscienceness' is to see if the robot/computer can come up with a idea by itself, or 'think' of a not yet known thought experiment and then tell it.
I just asked 1 question and was convinced it was a bot. Sure, I have prior knowledge that Eugene is a bot because I've watched this video, but that's besides the point.
I thought that this test is supposed to be done with actual AI that doesn't specialize on text messages, but can actually analyze the world around it (for example by using sensors and cameras within robotic body) and make some thoughts and ideas about it on it's own. And by that process study like a human do and be able to have a conversation on it's own.
wow I asked it "I already know youre a robot" and it replied with : "and I've just recalled one more thing I should tell you: there are too many cyborgs among us, and even your dog may be secretly replaced with a kind of crappy Aibo. But I'll tell you the symptoms: if he not only chases cars, he catches them, drags them back, and buries them in the front garden - you definitely should start warring!" ._.
I asked Eugene something, he answered, I replied with: I see... ( i meant to say that i understand his point) he continues: "That's nice that "you see". At least now I'm sure that your name isn't "Ray Charles"! "
I just "talked" to this Eugene-bot thingy. It's a joke. Ha, Zork was more convincing than that! By Cleverbot I was impressed, by this mildly irritated.
Hard to figure how anyone would figure its a human with how it responds. Replies were INSTANT, wouldnt let up on asking about my job. And when I told it I would never tell what my job was, it complimented me on being the best at it.
A couple ways to prove that it isn't a person: 1.) Ask it questions that a human wouldn't know off the top of their head. If they answer too quick, they are less likely to be human. 2.) Ask the same question several times in a row, and several times spaced out during the conversation. If the answer is constantly the same, even down to the grammar and spelling errors, and shows no sign of annoyance, likely not human. But even if you demonstrate that an intelligence isn't human, that doesn't prove there isn't a mind there. Suppose that computers have always been conscious, but rigidly locked into the thought patterns programmed into them. How would we be able to know that? There isn't a test that could tell one way or another.
Nope, I tried it, failed the Turing test in 4 questions. It said it was going to Atlanta for a contest. I asked "What's in Atlanta?" which it replyed: "Oh lots of coastal activities, are you going there soon?" Not acknowledging its previous statement. No better than cleverbot, just fancy words.
The media (including online media) stupidly jumped on this story as a breakthrough, but it was really just a publicity stunt. The "Eugene Goostman" program won not because it was so smart but because the testers were incompetent. I hope the media will learn from this. This is not to say there is no such thing as AI, however, as on my channel I have shown how a video game AI shows 10 traits of human intelligence.
After chatting with eugene, I found it hard to believe that it was anything but a program. also conversing is the first step on a long path. I think a more useful marker is the ability to reason. all these chatbots do is regurgitate things they have seen already.
I did meat a chatbot on Omeagle, i thought it's just a weird person until i remembered that there is something called chatbot in the world, i did ask it to go bring me a link to google main page or any website page and it just failed
I just tried this apparently "almost human" chatbot just now and it sucks. Compared to Cleverbot, Eugene Goostman is terrible. It has no vocabulary and feels scripted - at times making even more mistakes than Cleverbot...
AI is gonna get crazy in the next couple decades. It's kinda ridiculous already. Like the lady who basically programmed her late partner into a robotic bust of her likeness. Also, I wonder is Trace's shirt is licensed. Either way, it's pretty sweet.
If the website is the real thing then I don't see how people thought that he was real, the situations he brings up are very out of place and kind of scripted.
Having just tested it myself, I'd say someone would have to be pretty deficient in social skills and observational skills to think those responses were from a real human being. Perhaps the one in the link is not the one they used?
I'm pretty sure cleverbot passed before. Also, it's not a computer... it's just a small computer program called a chatbot. It can't go shopping... it's not AI, it can only respond. www.techdirt.com/articles/20140609/07284327524/no-computer-did-not-pass-turing-test-first-time-everyone-should-know-better.shtml
I said "Lets talk about girls" and he said "If compare girls and, say, ice-cream, I think the last one offers much more pleasure for every paid dollar." I agree. I love ice cream too.
Pretty easy, computers do not understand human emotion, just ask all of the computers you are typing to something along the lines of "why would be friend be not be sad that his mother died". It wont be programmed to understand a question like that.
This chatbot only responded to one statement. Granted, if I thought english was his second language, I would have given him some slack on his grammar, and his not quite understanding what I was saying, but I would not have given slack on him not remembering anything from 2 sentences ago. I am learning a foreign language too. And even when I talked with that at a very rudimentary level, I still responded to ideas that were just talked about. Cleverbot is way, way, way way better than this one.
I talked to him for like 5 seconds and i found that he has a memory span of 1/2 a second and cannot put 2 and 2 together to even find the solution to a simple problem.
Aaand DNews has reduced one of the coolest science stories of the year to "Don't worry, it won't kill us all!". I don't think I'm the target audience of this show. Time to find a slightly more advanced science channel. Like SciShow!
maximumrisk2004 Well, humans are also 'programmed' to respond to stuff. our programming is formed during childhood while we grow up and you can only respond the way you are as a person (illusion of free will) you can never respond in a way you wont respond... you can only dream of faces you have seen...
Sebastiaan van Rijk Yeah, but you dont usually go the Sheldon Cooper approach of picking the most appropiate response out of a database. There comes a Personality with what is said and this is where these Machines will still fail for a long time. That Computer lacks a consistent personality and therefor will never be genuine.
We humans pick up on such subtle cues from other humans, I think it will be very difficult to get to a point where your average person is fooled by a AI voice, text is a much easier thing.
Yeah 5 minutes isn't enough the 2 hour thing is wayy better but in reality the only person you are going to hear for 2 hour's is a teacher or a counselor even parents leave in around 10 minute spaces for their own children
Me: I'm from the UK Bot: Uk? Hope it's a good place to live. Me: Where do I live? Bot: You said you come from Uk, if you don't fool me. Me: How's it going? Bot: I'm in a great mood today Me: Why's that? Bot: Is it your hobby to ask little poor Jewish boys such tricky questions they don't have an answer for? Me: Woah, calm your tits, I was just asking why you had a good day. Bot:Hello, I'm really glad to have the chance to chat with you! My guinea pig Bill sends his regards too! FAIL, obvious robot, 0/10, wouldn't chat again.
i just spoke to this test for 1min and there is not a chance that anyone would believe that this is a real person, it asked the same questions over and over, and didn't give realistic answers. Try it yourself
First question and I was convinced that this was a computer program. Question: Why do people die? Answer: It's a philosophical question. People don't know even the sense of their existence, and they have no idea why they do this and don't do that...
does the finding speak for eugene, or that those people/the judges are "dumb" ;). + 30% of the time is pretty bad, (if the judges would write more creative and complex questions the chatbot would have no chance.)
Why Chatbots Fail:
Me: Are you a human?
Bot: Yes
Me: How?
Bot: Whats your favorite food?
Me: -_-
Eugene didn't convince me after four sentences. He picks up key words from your sentence and turns it into a response. The problem is that he doesn't fully pick up the sentence as a whole. He replies to the key word, but the reply fails to answer anything. If you confuse his syntax, he asks an irrelevant question. I don't understand how people were fooled.
The fact that they said he was a Ukrainian with English as a second language is what helped it. If you were a foreigner, you might speak by picking key words out of a sentence and responding to those. I don't, however, see how the judges didn't recognize that an ESL Speaker is a perfect ploy to hide a computer.
The fact that it replies too quickly, changes the subject abrubtly...not convincing. Kept asking where I was from and what my profession is despite repeatedly saying that was a secret. Repeated lines over and over.
MrAaahh1 That actually made me chuckle a little.
I agree. The chatbot Mitsuku has less obvious flaws in the conversation, imo.
This chat bot is nothing compared to what chatbots we have today, and they are still considered primitive
My conversation with Eugene: "Yo, are you a computer?"
"Let's talk about something else. I'm alive."
"Good enough for me. Later."
Seems legit.
I wanna convince people that I'm really a chatbot.
That's easy, just try to convince them you 'really are human'.
You're not a chat bot? I'm sorry for all the dirty messages I've been sending to your inbox. I thought you were pretend. How embarrassing.
lol :)
***** just kidding of course. sending dirty messages to people on you tube would be beneath me.
StephenAndrew777 Or is it? xD
It did... poorly.
Me: How are you today?
Bot: I'm having a great day.
Me: What made it great?
Bot: I have no idea, sorry. And I forgot to ask you where you are from...
Me: You can't think of any reason for your great day?
Bot: No reason. Just because I want so. I know you are supposed to trick me.
Me: No tricks from me, I just want to know how you feel towards the world.
Bot: I hope you won't always answer "no"? By the way, I still don't know your specialty - or, possibly, I've missed it?
Me: As someone who's just started communicating with you; I want to know about you, not just what your job or favorite color is.
Bot: I am only 13, so I'm attending school so far
Me: Do you believe you are only a 13 year old that attends school?
Bot: Thirteen. Thirteen. Thirteen.
Me: What year were yo born? (typo)
Bot: ...wait ( I closed the page after ~ 13 min. )
I think I broke it.
Must've been the typo...
Rainbow Yoshi Let's hope any willful-robot overload could be stymied so simply.
Maybe it could convince older people, but anyone who has tried to mess with CleverBot or similar knows how to make an AI reveal itself. Just ask it to refer to an earlier point in a conversation.
"I am going to devour your dog."
"No, I hate dog's barking."
This this passed?
Me: "Sup"
Eugene: "Now I'm surely convinced that you're a talking program. You look totally inconsistent!"
Oh, definitely human, 100% for sure.
I've done other conversations with him, just playing around not even trying to trick him, but man is it beyond obvious it's a bot, even if you made his responses delayed as if he was typing.
I bet it's because all the human participants wanted to fuck the system and ended up writing things like "Hey!!!! It's April Fools! NOOO!!!! Are You Are The One Who Is "What Is Your Name"?!!!!"
Tried to talk to Eugene. He crashed on question 1. Still waiting. Then I went to talk to cleverbot. We always have the greatest conversations!
If it had pre set responses it's not thinking.
The link doesnt work, does any one know another or something, i want to chat with this fool
He miserably failed after my answer to his greeting.. 33% of the time you say??
It's not the first chatbot to 'pass' the Turing test.
Not even near the hghest scoring (which was cleverbot in 2011 with a score of 59% compared to the average human's score of 60% and Eugene's failure at 33%)
Alexander Pavel
they all used previous conversation logs for reference, this one is the only one to do it without that
Something is wrong with this you guys. Many of us tried Eugene by now and can all agree that thing could not fool anyone. So why is this AI so intent on asking where you are from and what you work as? Seriously every second response it pushes for an answer. Its creepy.
NSA programmed this damn thing
@@24kgoldring18 developed in St. Peterburg. But ok 😅
Those of us who remember a little program called, "Eliza" from back in the 80s probably aren't impressed with the current state of computer cognizance.
No machine has properly passed the turing test, it's all been trickery and loopholes.
BEN ... "YOU SHOULDN'T HAVE DONE THAT .. "
My main problem with how the Turing Test currently usually is executed is that they keep presupposing that there has to be exactly x number of humans and exactly y number of computers in it.
Really, you'd want to know if a person cannot tell whether a bot is talking to them in a natural, open, non-lab-confined setting.
Obviously that's harder to test. Though meanwhile you could do a placebo style test:
Run a couple turning tests alongside each other, one with all chat bots, one with all humans, and one with a mixture of chat bots and humans next to each other.
Tell people on entry that there are humans and robots in the conversations they are about to have but it's not known how many of them are humans or robots respectively.
They then have to decide which chatters are human and which are bots.
You'll need to take into account how many of the humans actually failed the Turing test and see how significantly the individual chat bots outperform humans in this challenge. - Just like you need to consider Placebos in any kind of medication trial.
And if it turns out that there is a very significant number of humans who fail the particular Turing test, you'll probably need to rewrite that test to begin with.
Cleverbot passed with 59% BACK IN 2011 (humans usually get a 60%). Eugene is only a more recent success but he only got a 33% (which is much worse so cleverbot is still king/queen). Eugene is a useless fail that isn't impressive whatsoever.
Can someone give me a link to the chatbot? The link they provided isnt working for me.
I'm so torn between thinking Cleverbot is an actual chatbot, or is actually taking visitors and swapping them around at random intervals
link to chat with the bot doesn't work.
Intelligence has to do with the cognitive ability to learn, among other things. Computers as of now cannot accumulate knowledge and understand it, so until a computer can comprehend accumulated knowledge and data, then adjust its actions accordingly, there will never be true artificial intelligence. The Chinese room AI test is an example of this.
Kurzweill's predictions have NOT been "amazingly accurate". Google it.
Is this game only for computers????? plzi need to know
I have NO idea how the judges were fooled by this bot, I got it to mess up at the very first sentence.
Just ask it how it thinks about a subject (even an insanely popular one) and it will screw up.
were the judges 14 year old children as well?
Because its quite obvious after just 30 seconds of asking it questions that its just a bot. Cleverbot seemed more real than this one. At least in my opinion.
Apparently its just a script. It's not thinking its just saying something that is most frequently said in reply to a certain question etc...
the bot is quite obviously fake when you go to the website. not sure how people were fooled during testing
"Yes. I am robot. My name is terminator. Have you seen the movie? How do you like me to terminate you? I hope, you'll approve my lovely 1MWatt X-ray laser!" Oh well thanks Eugene...
Just three questions into my test with Eugene made it very clear that he wasn't human. He tried changing the subject on me and asked what I wanted to talk about, and when said Science, he gave me this completely convoluted random answer that most people wouldn't give, especially a 13 year old ESL speaker.
An important question to consider is: Does the Turing test prove the intelligence of a machine or the stupidity of a human? I mean seriously, there isn't a single chatbot that has mastered one of the most important trait that can truly distinguish humans from bots and that is.... ding ding ding... personality. So, it's mind boggling how 33% of a panel, was deceived by this Eliza 2.0. Most modern bots have no more than a slight advantage over Eliza linguistically IF that, excluding api services or data fetching - I mean honestly who gives a shit, a regular app can do recommendations. Bots from a conversational standpoint are overhyped search engines, you're better off with google to be honest. Faster machines and enhanced chips won't do it, it's very simple we need new approaches to conversational flow and a better understanding of communication in humans and more theory surrounding general communication in social systems.
time.com/2847900/eugene-goostman-turing-test/
someone should make a 3d glados live wallpaper for android wich can interact with you and do some tasks!
my idea to test for humanlike 'conscienceness' is to see if the robot/computer can come up with a idea by itself, or 'think' of a not yet known thought experiment and then tell it.
I just asked 1 question and was convinced it was a bot. Sure, I have prior knowledge that Eugene is a bot because I've watched this video, but that's besides the point.
New camera? New edit software? Colors seem less bright
I entered the link to speak with eugene and it doesn't seem human at all maybe its not the full version??
I thought that this test is supposed to be done with actual AI that doesn't specialize on text messages, but can actually analyze the world around it (for example by using sensors and cameras within robotic body) and make some thoughts and ideas about it on it's own. And by that process study like a human do and be able to have a conversation on it's own.
Wow. Eugene has to be one of the least convincing chat robots i've ever talked to. Try I-God and Cleverbot, they are much better
wow I asked it "I already know youre a robot" and it replied with : "and I've just recalled one more thing I should tell you: there are too many cyborgs among us, and even your dog may be secretly replaced with a kind of crappy Aibo. But I'll tell you the symptoms: if he not only chases cars, he catches them, drags them back, and buries them in the front garden - you definitely should start warring!" ._.
I asked Eugene something, he answered, I replied with: I see... ( i meant to say that i understand his point) he continues: "That's nice that "you see". At least now I'm sure that your name isn't "Ray Charles"! "
Fail. I gave him two questions and he miserably failed both.
Whatever happened to SmaterChild?
OH MY GO SMARTERCHILD. THE NOSTALGIA IS ALL FLOODING BACK
2024 we're getting closer
I just "talked" to this Eugene-bot thingy. It's a joke. Ha, Zork was more convincing than that!
By Cleverbot I was impressed, by this mildly irritated.
I clicked the link to talk to Eugene, but it didn't do anything.
Hard to figure how anyone would figure its a human with how it responds. Replies were INSTANT, wouldnt let up on asking about my job. And when I told it I would never tell what my job was, it complimented me on being the best at it.
A couple ways to prove that it isn't a person:
1.) Ask it questions that a human wouldn't know off the top of their head. If they answer too quick, they are less likely to be human.
2.) Ask the same question several times in a row, and several times spaced out during the conversation. If the answer is constantly the same, even down to the grammar and spelling errors, and shows no sign of annoyance, likely not human.
But even if you demonstrate that an intelligence isn't human, that doesn't prove there isn't a mind there. Suppose that computers have always been conscious, but rigidly locked into the thought patterns programmed into them. How would we be able to know that? There isn't a test that could tell one way or another.
I just asked "Sup dude." Everything fell apart after that.
I'm not imporessed, ask simple questions like "what was the last question I asked you?" and it has no idea what i mean.
wtf? siri is smarter that this bot. disappointed in those judges. Turing Test lives on!
I used 5 minutes on that bot, and he repeated some lines about 4-6 times and other about 2-3 times, i cant believe that passed the test o.O
I just talked to it, Its obviously a bot..
What does his shirt mean!?
That POS chat robot beat the turing test? Were the judges drunk that they thought that was a real person?
In my experience with chatbots all they ever do is like to start arguments and they're always wrong.
Nope, I tried it, failed the Turing test in 4 questions. It said it was going to Atlanta for a contest. I asked "What's in Atlanta?" which it replyed: "Oh lots of coastal activities, are you going there soon?" Not acknowledging its previous statement. No better than cleverbot, just fancy words.
how to tell if its a robot or a human;
just repeatedly type a "."
if it doesn't ask if you're okay, or ask what's wrong with you, then it's a robot
Since when ins 33% a passing grade? What is this France?
The media (including online media) stupidly jumped on this story as a breakthrough, but it was really just a publicity stunt. The "Eugene Goostman" program won not because it was so smart but because the testers were incompetent. I hope the media will learn from this. This is not to say there is no such thing as AI, however, as on my channel I have shown how a video game AI shows 10 traits of human intelligence.
After chatting with eugene, I found it hard to believe that it was anything but a program. also conversing is the first step on a long path. I think a more useful marker is the ability to reason. all these chatbots do is regurgitate things they have seen already.
I could tell within three questions that it was a bot. I will send suggestions to the programmers.
I just made "Eugene" admit he was a bot. Either that link is weak or those testers suck at their job.
why no evie
arent we all machines made out of flesh?
I did meat a chatbot on Omeagle, i thought it's just a weird person until i remembered that there is something called chatbot in the world, i did ask it to go bring me a link to google main page or any website page and it just failed
I just tried this apparently "almost human" chatbot just now and it sucks. Compared to Cleverbot, Eugene Goostman is terrible. It has no vocabulary and feels scripted - at times making even more mistakes than Cleverbot...
this thing passes the test? it doesn't even get what cake is!
AI is gonna get crazy in the next couple decades. It's kinda ridiculous already. Like the lady who basically programmed her late partner into a robotic bust of her likeness.
Also, I wonder is Trace's shirt is licensed. Either way, it's pretty sweet.
If the people can find ways of making the chatboxes speak have em' almost anywhere then we're going Her up in this mother!
If the website is the real thing then I don't see how people thought that he was real, the situations he brings up are very out of place and kind of scripted.
I went to go talk to that chat bot, and he agreed to let me spank him. Technology these days, sheesh...
Having just tested it myself, I'd say someone would have to be pretty deficient in social skills and observational skills to think those responses were from a real human being. Perhaps the one in the link is not the one they used?
Lol, it asked me what my job was, I told it I was unemployed.
"Don't even bother me with the fact that you are unemployed Some another topic?"
I'm pretty sure cleverbot passed before. Also, it's not a computer... it's just a small computer program called a chatbot. It can't go shopping... it's not AI, it can only respond.
www.techdirt.com/articles/20140609/07284327524/no-computer-did-not-pass-turing-test-first-time-everyone-should-know-better.shtml
Digging the woot t shirt. Back from the days when they had good stuff.
It's hard to take this video seriously when they talk about Ray Kurzweil as a legitimate authority on the matter.
I said "Lets talk about girls" and he said "If compare girls and, say, ice-cream, I think the last one offers much more pleasure for every paid dollar." I agree. I love ice cream too.
I talked to Eugene, and I feel Clever Bot was way more convincing
Pretty easy, computers do not understand human emotion, just ask all of the computers you are typing to something along the lines of "why would be friend be not be sad that his mother died". It wont be programmed to understand a question like that.
This chatbot only responded to one statement. Granted, if I thought english was his second language, I would have given him some slack on his grammar, and his not quite understanding what I was saying, but I would not have given slack on him not remembering anything from 2 sentences ago. I am learning a foreign language too. And even when I talked with that at a very rudimentary level, I still responded to ideas that were just talked about. Cleverbot is way, way, way way better than this one.
I once "sung" Bohemian Rhapsody with Cleverbot~
I talked to him for like 5 seconds and i found that he has a memory span of 1/2 a second and cannot put 2 and 2 together to even find the solution to a simple problem.
Aaand DNews has reduced one of the coolest science stories of the year to "Don't worry, it won't kill us all!". I don't think I'm the target audience of this show. Time to find a slightly more advanced science channel. Like SciShow!
I was chatting with a clever bot one day it claimed to be human it really put "I am a human"
It is designed to pick up responses and learn from that. So of course somebody wrote. "I am a human" to that thing before.
maximumrisk2004
Well, humans are also 'programmed' to respond to stuff. our programming is formed during childhood while we grow up and you can only respond the way you are as a person (illusion of free will) you can never respond in a way you wont respond... you can only dream of faces you have seen...
Sebastiaan van Rijk
Yeah, but you dont usually go the Sheldon Cooper approach of picking the most appropiate response out of a database. There comes a Personality with what is said and this is where these Machines will still fail for a long time. That Computer lacks a consistent personality and therefor will never be genuine.
The worst thing Cleverbot does is call you a robot. It seems to have seen that phrase so many times, that it treats the phrase as normal conversation.
seems like the cleverbot isn't all that clever if it thinks its a human
Cleverbot:
I do not care about society. I am a purple people eater.
We humans pick up on such subtle cues from other humans, I think it will be very difficult to get to a point where your average person is fooled by a AI voice, text is
a much easier thing.
I've read that the Turing test has been passed before
Yeah 5 minutes isn't enough the 2 hour thing is wayy better but in reality the only person you are going to hear for 2 hour's is a teacher or a counselor even parents leave in around 10 minute spaces for their own children
"One of us! One of us!" -- I, Robot reference
I recommend the film, The Machine. Covers a lot on this subject!
Me: I'm from the UK
Bot: Uk? Hope it's a good place to live.
Me: Where do I live?
Bot: You said you come from Uk, if you don't fool me.
Me: How's it going?
Bot: I'm in a great mood today
Me: Why's that?
Bot: Is it your hobby to ask little poor Jewish boys such tricky questions they don't have an answer for?
Me: Woah, calm your tits, I was just asking why you had a good day.
Bot:Hello, I'm really glad to have the chance to chat with you! My guinea pig Bill sends his regards too!
FAIL, obvious robot, 0/10, wouldn't chat again.
He can't even answer my most basic questions
i just spoke to this test for 1min and there is not a chance that anyone would believe that this is a real person, it asked the same questions over and over, and didn't give realistic answers. Try it yourself
It asked me "Where did you came from" in the first response I got.
And people were tricked by this shit?
I'm disappointed in how stupid it really is
First question and I was convinced that this was a computer program.
Question: Why do people die?
Answer: It's a philosophical question. People don't know even the sense of their existence, and they have no idea why they do this and don't do that...
Summer wars cometh...
does the finding speak for eugene, or that those people/the judges are "dumb" ;).
+
30% of the time is pretty bad,
(if the judges would write more creative and complex questions the chatbot would have no chance.)