3 PROVEN METHODS TO FOOL AI CONTENT DETECTORS | How to Trick AI Content Detectors
Вставка
- Опубліковано 19 січ 2023
- So you want to know *how to trick AI content detectors*. You, a human, want to figure out how to disguise robot-written writing from another robot.
This is something I've been thinking about a lot lately, and I found three ways to do it.
I want to tell you upfront: if you want to know how to trick AI content detectors JUST to sort of scam or cheat your way into creating garbage content, this video is not for you.
This video is for people who want to know how to trick AI content detectors for GOOD reasons. For ethical ones. Just so you don't get caught in the crossfire of the AI content detectors that are about to crop up.
If you're using ChatGPT for research and want to make sure your writing doesn't set off any AI content detectors by accident, these three methods may help you understand how to trick AI content detectors.
THANK YOU TO:
- Epidemic sound for the Music
- Pexels for the videos
This video is about: how to trick AI content detectors, ai content detection tools, how to bypass ai detection, how to trick an AI, how to beat ai content detectors, how to use chatgpt, how to tell if something was written by AI, how to use AI text generators - Наука та технологія
Your suggestion to simply write stuff yourself is RIGHT ON. Students at my college are spending so much time rewriting AI generated text and using tricks to fool AI detectors and it's actually less time consuming to simply write it yourself.
if you have adhd and hate that subject or the fact you need to write to much. it takes way less time to proof read and adapt ai text. oh and being dyslexic doesnt help at all either at writing it yourself.
the secret lays in prompt engineering. feed the AI some info before you let it write a text. for instance:
"When it comes to writing content, two factors are crucial, "perplexity" and "burstiness." Perplexity measures the complexity of text. Separately, burstiness compares the variations of sentences. Humans tend to write with greater burstiness, for example, with some longer or complex sentences alongside shorter ones. AI sentences tend to be more uniform. Therefore, when translating and writing the following content I am going to ask you to create, I need it to have a good amount of perplexity and burstiness. Do you understand?"
"using the concepts written previously, now act as an [fill in] and [perform action]"
this way, I get results that are not easily detected.
Or... ask your teachers for accomodations due to your learning difference. The current school system is bad for learning, especially with learning differences, such as getting bored easily. A single learning method (i.e. just reading, or just listening) is bad for learning, because your brain needs multiple stimuli to remember well. Also in many countries, chuldren spend more time in school, and have start times way too early for child or teen natural sleep schedules. Things like ADHD are also definitely overdiagnosed, due to lackluster companies seeking drug money, and the symptoms being prevalent due to the increase in games, mobiles, and the internet.
Just because learning is hard or a struggle doesn't mean we should reduce the effort to 10%. If you don't fully understand it, you will struggle at learning more. Yes learning shouldn't be focused on pure writing, and in most cases it isn't. Maybe it will cause a shift in stubborn systems, but they really need overhualing.
From a reader's perspective, it doesn't matter if the content was written by a human or an AI, as long as they enjoy the text. Aaaand, that's a problem.
That's the issue, ai text reads like, erm, ai text. It is often full of complete garbage,. Chatgpt was trained on 500gb of data. My hard drive is bigger than that. It is many decades away from being able to write proper human content, unique thoughts or anything other than a garbled rewrite of the top few results of google. Which incidentally is very close to stolen content in many cases I have observed with chat gpt3.
@@trading-university. No, it was trained on 60TB. Here's the info from ChatGPT "As an OpenAI language model, I was trained on a very large dataset of text called the Common Crawl. The Common Crawl is a repository of web pages and other online content, which currently contains over 60 terabytes of data in multiple languages. The exact amount of data that I was trained on is not publicly disclosed, but it is estimated to be on the order of hundreds of billions of tokens (words and other symbols). This massive amount of data helps me to generate human-like responses to a wide variety of questions and prompts."
aint that the truth. but Academia doesnt want that
When its wrong
Except you never sampled every single book, much of the internet, and compressed it all into
Hi Zulie, great video. I'm just beginning to learn more about ChatGPT, and I'm trying to determine what all the hype is all about. You did a great job explaining the service's pros and cons. In the end, everyone will have to use their own imagination and knowledge and write original content themselves.
I just copied and pasted a text written by a Master Degree psychologist into an AI Content Writer Detector and it only gave it a 5% chance of being written by a human. The detector determines if it's AI-generated based on the intelligence of the writing, the smarter it is, the less likely it's a real person! Haha😂
oh no hahaha. hopefully the detectors get smarter at picking out the real smart people from the smart AI!
@@ZulieWrites That was literally my thought when Chat GPT gave the reasons for its decision. I was like, "Soooo... anyone who can speak intelligently about their subject with proper term use and organization, knows proper (insert language here), and edits their work before posting is going to get flagged?"
Talk about flipping the script. That is so funny.
Great advice. This is one of the best videos I have seen about this topic.
Amazing approach, the testing method is incredible, many Thanks 😊
Awesome one, Zulie! You nailed it.
Many professionals, from various industries, are dreading losing their livelihood to AI. What they fail to realize is that it's inevitable and WILL happen, sooner or later. The best approach would be to embrace the change and equip oneself to involve AI in their profession and live in harmony. In my humble opinion, a person who doesn't embraces AI would ultimately benleft behind. AI won't take over your job. Instead, the person who accepts the inevitable involvement of AI in our lives, in times to come, would definitely push you out of whatever your profession is.
Wait a minute. I have a Journalism degree, and everything I produce is concise and unique with no spelling, grammar or punctuation errors. Am I to understand a bunch of hack AI detector programmers would call me a fraud because the quality is too good? Nothing I write is boring, but "boring" absolutely is not an indicator of AI content. There are plenty of topics that are boring, but about which still need to be written.
Beautiful video and message 💙
I think it’s funnier that chatgpt thinks human writing is terrible in comparison to AI and you think it’s the other way around that creates garbage content. I think it is possible that with like version four and five and six that it will be able to 100% mimic us. This stuff is here to stay. It didn’t take you much to fool artificial intelligence checkers but when you wrote your own story out, It got a worse score. I think the best writing is somewhere in between Ai and human text.
I don't use AI-generated *content* in my writing (a sci-fi novel). Instead, I use AI to proofread my work, and I ask it to offer suggestions on cleaning up some of my text, whether adding a bit more description to make something clearer or cutting down on text and making it concise where it needs to be. I also use it to check for active voice, present tense, and whether or not there's a balance of show and tell. Overall, I do all the actual work.
I think people should embrace AI as a tool to help make their jobs easier.
Here's what an AI thinks of my writing so far, despite using AI to help me clean it up a lot.
Me: "From what you know of my writing with your updates, does it read like a person or an AI wrote it?"
AI: "Based on the writing you've provided and the guidance you've given regarding adverbs, spelling, and prose, the text reads more like it was written by a person. It exhibits a natural flow of dialogue, character interactions, and narrative elements, which aligns with the style and preferences you've conveyed. It doesn't have the rigid or formulaic structure often associated with AI-generated text."
Thanks for the video.
if you put the mic on the other side of the brush, do the bristles act as a pop filter?
Great video. All the writers' community is pissed off by the use of AI to make crappy content by some people. It seems you have made the content just to address this issue and tell those people to apply their brains and write something creative.
Great video! Thanks!
Very good advice!👍👍
I’ve tried Novel AI but the reality is AI can’t really write with the same style that I want in my writing. AI content is ok for writing simple things like captions, but for short stories and novels it’s pretty useless
So be real. Use your skill. Learn more. Work hard. Make ur goal success for real if you want🤷♀️
Nice video, enjoyed it.
So "making mistakes is the low-hanging fruit" is quite true. Even "word" corrects your mistakes. Having chat write in a conversational tone is a small step forward. Tossing in an anecdote is a large step forward. I enjoyed the video. However. I think you could have used a more upscale content detector. This is a war, and you must fight fire with fire. Thank you for the video it is quite helpful for those who do not want to pay or subscribe to AI services...
I think that Content at Scale has changed some things. I tested its capabilities with essays that I wrote before these AI programs were open to the public. Content at Scale is flagging these as "likely both AI and human". I attempted to rewrite the sentences and paragraphs that it is flagging with no success. I even asked chatgpt to rewrite them as a human with no success.
Can we use articles we have published in magazines or other media on Medium and after we have permission to reprint? Do we need to give an attribution notation on the article?
You absolutely can as long as those other places are ok with it! It's best practice to say where it was originally published.
Awesome! 😀
I find it more satisfying to write a content myself and having it ranked than the other way around. Ai is just a brain storming tool nothing more. Can’t replace me yet:)
Give it time.
Hi Zulie! Thanks for the video. I think that soon it's going to be a marketing strategy to say that something was created by a human not an AI, just to set it apart from ordinary content. It'll be what carefully handmade products are when compared with mass-produced ones. There will be those who won't care either way, but there'll always be those who appreciate "human made" things.
I love this thank you so much!!! This was the response I got from ChatGPT Me: Can you tell if this was written by AI? ChatGPT: As an AI language model, I don't have the capability to directly determine if a piece of text was written by AI or a human. However, the given text could have been written by either. It contains a well-crafted marketing message, emphasizing the author's abilities to create engaging content with a mix of storytelling and SEO knowledge to help brands stand out and make a lasting impact on their audience.
In general, it has become increasingly difficult to distinguish between AI-generated content and human-written content, as AI language models have become more sophisticated in mimicking human writing styles and producing natural-sounding text. Nonetheless, the content provided seems to be in line with the type of text that an AI language model could generate for marketing purposes.
interesting video, thanks for sharing!
Thank you.that was good
The first time I have seen someone giving great advice. What you said about writing yourself is absolutely true! Thank you very much!
Nicely done, as usual, Zulie. Thanks for experimenting. I suppose AI can be marginally helpful as a "jumping off point" or research tool, but I still don't plan to use it. At all. Doing my own research, writing my own words, and telling my own anecdotes, while often time-consuming, is still the best way to produce a quality article. Months or years from now, if an AI content detector wrongfully accuses us of being robots, well, I guess we'll have to write more informally. Add more stories. Something. Thanks again!
This is going to be a constant battle forever between AI writing and AI detectors. The better AI writers get, the harder it will be to differentiate them. I mean, I can have AI write a whole article about a specific topic, and trust me... it does a much better job than I do. (I am no slouch, but I don't have any Pulitzer prizes either). So the questions then becomes when does the snake finish eating its own tail, and when do AI detectors start detecting legitimately written human content as AI content. Quite the catch-22.
The problem with AI detectors, unfortunately even the ones that rank your content, is shown here in the video. ChatGPT said it could tell it was AI content cause it was well written, technical, and with few errors. And, that the human one was detected because it was more conversational and had errors (plus personal anecdotes).
So basically, if you write a well written article in a technical tone, it'll get flagged as AI content. Thats kinda f***ked up😆
I just watched your video again. Your presentation is terrific and the summation was icing. I could not agree more when you stated that you need to write SOMETHING! There is no need to be a lazy ass...thanks again
Glad you liked it!
Well Said Zulie Rane, Human Content is Original always
Eeeeexcellent video! I agree 100%. My background is in PR/Marketing and I have degrees in English, Biz Dev and Psych. AI will not replace humans, it will only replace humans who don't want to learn. But it will also not allow humans who don't know what I do to replace me. The human element, talent, wisdom will always trump AI or anyone trying to use it to cheat. However it is a great tool and can be a way to close gaps on unfair barriers to entry (like simply not being able to hiring assistants). Great job!!!!! 🥰
Wow... Didn't think of putting it back in to Chat GPT to help solve the issue... "The call came from inside the house".
Nice, it worked for me☺
I was going to say chat gpt probably remembered writing it since it remembers previous conversation..... Ask GPT to write it in your tone of voice and I would love to know what you think. You could just ask chat gpt to write another paragraph for you.
Thank you!
Will the contents of articles summarized by AI tools with Scispace also detected and considered PLAGIARISM? Pls share me your knowledge about it.
I have not heard of that tool!
I don't think is about what they're not the person can write I think it's about how much time does it take. I've been riding since I was 7 years old in the last people ask me what university I was in or go to. Same thing with drawing. If you want to be creative that's fine if I'm going to make money and trying to get it done as quickly as possible a lot of cases has nothing to do with whether I can or cannot do it.
Pretty you should paste at 7:23 the text without double quotes "" then it should gives you the text is AI, check it out
I hate it when I write the whole thing and AI keeps changing it sound, when I only ask it to check clarity and grammar.
Include “in human language” in the prompt. Better to usecAi to generate research and idea variations, then use that to write your own content.
But zerogpt even detects that(remeber there is a difference between gptzero and zerogpt?
Factual content gets flagged as AI. What should I do if I am writing something historical? I think I may have to be excessively flowery in my writing and use a ton of personal accounts which is not fair, and often inappropriate. The ultimate advice in the video is an oversimplification.
Thanks zulie 🥰
So am I led to believe that if an entire piece of text was written by a human , that all AI generators would and could never say ,it as AI generated?
i really love your videos i think a.i tools like gpt-3 ae not tools make new content they are basically tools to minimize the time it take you to research on a certain a topic think of it as a advanced google when you search something on google it gives you every result internet has to offer on your research and rank it on the basis of seo but gpt-3 kind of like a search engine on which you can search anything and it will give you the best result on your topic based upon its gigantic memory and things it have learned soo it is merely an efficient way to search stuff thats all
But dont think a.i cant get better even tho gpt-3 is just an efficient search engine the time is not far when a.i becomes really powerful and becomes efficient human beings.
I mean after all thats how we have been evolving from monkeys to cave men to empires to modern day humans we as today are in a sense an optimized version of our ancestors cave mens soo dont you think in few decades a.i is the next evolution in humans?
Great , the main thing is that the time you are wasting in converting AI generated contents to human generated so that should not be detected , invest that time in searching and writing your own, use AI as a tool , as a helper
What if we human write just like an AI(professionally)? AI detectors sucks as long as long as written content is enjoyable and well appreciated by others. If detectors are trained for detecting AI content, then we could also trained our kindness/mind to write like them.
The dirty little secret to fool AI content detectors….WRITE! 😂 It’s totally true. Good writing is fool proof.
The last advice is the best!!! You don't want to start building your career generating your writing with AI. Eventually maybe not now but in 3 or 5 or maybe 10 years a tool might come out that can check for the originality of your past work and it come out that is all generated by AI. Yes its faster to use AI but in my opinion using AI to actually learn how to write better is a much better way to use AI. And I think AI is an AMAZING tool that can still help us so much!!
Why should you be not allowed to use AI for generating content?
Isn’t it silly on one side we make it really easy to develop content using AI and on the other side we block the content generated by AI.
Its like driving a car with the brakes on 🤓
Well when I pasted one of my poems and asked Chatgpt to tell whether it's written by an AI or not ?
You know what it said yes.. I wrote that poem back in 2019 when ChatGpt was not even a Thing!
LoL.. Now what?
The way some news articles are written they should be using AI.
Bingo!
You sound a bit hypocritical! I understand what you're saying. I don't know about you, but I have written plenty of blog posts for companies where I have to write using their tone and voice and my name is given as credit for writing the article all the time. The same goes when I am hired for ghostwriting. Sure, when we write on Medium we can use personal anecdotes, but I can't do that when writing for a professional-client when the writing has to be more formal and maybe "boring" to the average reader. At any rate, this is a good video. Thanks for sharing.
Kindly tell me how I will withdraw my earnings from medium?
Born on third base due to looks and thinks she hit a triple. AI will destroy most writing gigs anyway.
I get so frustrated with AI content detection because half the time it says articles I wrote myself are AI generated.
Thank you for this - I'm only a month or so into exploring AI (mainly ChatGPT) and I've pretty much ended up at the same place. I'm finding it a great tool to use in the middle of my workflow:
1: Identify personal & original content and write it down as a braindump
2: Re-order into some logical construct
3: Enter into ChatCPT to create text in a comprehensible format
4: Rewrite in my style. This usually involves a degree of profanity that puts ChatGPT into meltdown 🤣
It seems to be a great productivity accelerator or a research information regurgitator but any originality has to come from me.
Thanks again
No I tried asking chat gpt to rewrite content in a more humanized manner. And the detector still flags it as AI
Yes that's me
And my assignment is due tomorrow. If even I chnage the wording
what if I'm trying to sound sophisticated and use complex vocabulary
Haha then as long as you provide original insights, you're golden!
Good job! The word today is stragety. Not willie nillie. Think folks and listen to her!
I do not agree with you concerning the robotic output of AI machines like ChatGPT. Already these machines are producing content that, although robotic in nature at this moment in time, is highly readable and makes sense. This is AI in its infancy. Not long from now it will be able to produce content with all the personal human touches that will make it impossible to tell the difference between human and machine generated content. Machine produced content "with personality" is just a couple of years away and once that happens AI detectors will become redundant. I am no fan of AI by the way, I can't help but think that what is a useful tool today, will be my boss tomorrow. But maybe I am just a pessimist....
Love u
All new technologies are dammed as bad and problematic when they enter the mainstream. It’s here, and people will have to adapt, even professors.
i cant hear u , why so monolithic
Google said AI is acceptable. You dont get penalized.
You still get penalized for "low quality content" which, atm, 100% of pure AI gen is. www.theregister.com/2023/02/09/google_reverses_ai_page_spam/
Informative video but background music is too loud and chaotic. I literally had to pause the video to make sure that that music was actually coming from the video. It was especially distracting when using headphones.
Thanks for the feedback! I've heard this from a few folks, so in my newer videos I'm testing out no music except on title changes.
If there is nothing wrong with what you are doing then why do you need to use a "trick" at all? Answer: Because there is, at least to google, something wrong with what you're doing.
ZeroGpt detect ai content 😢
This is entirely illogical. AI says us that you cannot write perfect sentences as me? Why? You are trained on the articles and books written by human.
*[Title] How to: Trick AI Content Detectors*
Write yourself the content???
Pretty much!
3 Proven methods to fool AI content detectors and then "write by yourself" What the holy cow? So, you already trick most of the viewers. Be honest with your title.
speak louderrrr
zulie, yesterday I finally reach 100 follower, and im so glad about it. So so glad about it because I thought my hardwork research etc come to fruition.
No joy can compare to this feeling because for the first time I feel like I belong, writing is my passion and it save me from the darkest part of my life.
Turns out MPP program put me in different feelings, it really broke my heart and soul.
what do you think I should do ? Do you have any kind of invo that I might help me ?
it really hurts me so much, because payoner, paypal, and stripe works in my country. but somehow they still prohibited me to join.
it so unfair 🙁☹️
That helps me blog with happiness.