How To Detect If Chat GPT Wrote An Essay (Plagiarism Checker) - GPT Zero

In this video, the speaker is introducing GP2 zero, a new tool designed to detect and protect against plagiarism and academic dishonesty, specifically when it comes to using Chat GPT. The speaker demonstrates how to use the tool by having Chat GPT (a language model) write a 3000 word essay on the topic of whether or not Hamlet loves Ophelia.

The speaker then copies and pastes the text into GP2 zero, which then checks for perplexity and bursting. Perplexity is a measure of how well a language model is able to predict the next word in a sequence and the lower the perplexity, the more likely the text was generated by a language model. Bursting checks for repeated phrases and the higher the bursting, the more likely the text is generated by a language model.

In this case, the tool indicates that the text has a low perplexity and a high burst, indicating that it is likely to have been generated by a language model. The speaker advises that while GP2 zero can be useful as a tool to detect plagiarism and academic dishonesty, it is not a substitute for independent thought and writing.

#gptzero #chatgptplagiarism #plagiarismdetector

[luxvincunt]

I got in trouble for "ai plagiarism" today even though I wrote my essay myself :(

[MNVV]

I have discussed it with my professor who had used it for my discussion board assignments. He checked multiple times saying that my discussions where AI generated and that some times they arent. He checked with other students in my course also having the results being mixed as well even though we had written them ourselves. Of course he did say that he won’t rely on it unless he finds repeated patterns such as formal or professional writings, in my case were not. my discussion board threads were not repeated and were answered in a neutral tone, since I try not to be one sided but I still had to prove my point and I used sources do to so.

[drit]

Just scanned my final paper from 2019 in GTP zero and it says it's 0.5% written by human! Damn, I basically did chat GTP job even before it was invented... I'm that good I suppose!

[jhtompkins]

Edward (Tian), I hope you see this. I'm a JC English prof (and Princeton alum) in Los Angeles and I'm writing you here because I couldn't find your contact info at the GPTzero site. My wife and I are both Eng profs and ChatGPT is causing us headaches (it's availability has spread thru our students like wildfire). Speaking as an English instructor I hope to avoid spending most of the semester trying to figure out who's plagiarising or using AI-generated text in their essays. I believe strongly that teaching reading and writing is worthwhile - on the somewhat remedial levels that we find in the JC world, but also in more rarefied academic settings. At this point, knowing my students, most of them will give chatGPT a whirl during the spring semester. If nothing else (and there many things that could be lost to AI) it means that students won't read the books we've assigned. I take that loss personally. Does it matter that my students read "Beloved"? "Kindred"? "The Illiad"? "Faithful Place"? "Generation Zero"? "The Long Ranger and Tonto Fistfight in Heaven"? It's toughl to generate empirical data proving the impact of this AI workaround (not yours, I mean the original). But I generally love the books I teach, and losing the leverage I have in trying to interest my often uninterested students in reading feels somewhat like not a death in the family, but maybe a serious illness (after all, the bottom line, to put it bluntly is this: read or fail). Anyway, contact me and I'll keep you abreast of my experience this spring. I have great hope that GPTZero can help keep ChatGPT from overwhelming my efforts to keep kids reading. Thanks, and Cheers!

[alijamaslesa]

This tool has a long way to go. I have plugged in multiple essay's and discussions I've written 100% in my own words, and for every single one it has told me it was written by A.I. Both ChatGPT and ZeroGPT are going to destroy students. If I was a professor/instructor I wouldn't solely utilize ZeroGPT, because now you're going to falsely accuse students who have done nothing wrong. I also did an expirment where I used different Paraphrasing tools on everything I have written and ZeroGPT most of the time doesn't detect it was A.I written, completely odd that it accuses you of A.I writing when it's in your own words, but using a A.I parapharser it doesn't. The gentelemen that created this did not do the appropriate amount of work on it, seems that he just went out his way to create something quick to grab headlines and fame.

[argophontes]

You can ask ChatGPT to write the essay in a way that is difficult to detect by AI detectors and the check will come back "Written by a human". If universities really start to implement this, especially knowing that OpenAI's checker for their own AI gives a false "Written by AI" result at least 9% of the time, this could set a really dangerous precedent in academia.

[truthfullight]

Why not test it with human written essay?

[sobe]

My essay says it was generate entirely by chat gpt even though I did not use it for the essay. How can you help?

[dannyinfinity]

when I'm trying to find out if I plagiarized it doesn't necessary detect it because I used Grammarly to help me out with grammar and when I upload my file to Grammarly plagiarism it does detect plagiarism but idk where or how bad is it. I don't want to pay premium. If its few percentages of plagiarism, will I still be safe or would I be in trouble and I have to explain my professor about it see what she says.

[Jrush]

I still think it could be avoided if you asked it to write an essay or story with little to none perplexity and complexity. I tested it and it said that it is still likely to be written by a human based on the perplexity score.

[mainaksarkar]

hi sir, just to point out, the student who made that need to check his algorithm, I have tested the model, there, the lines which i have not copied from chat gpt and even have not taken the idea, is giving the message that, I have copied the lines from AI and the probability is 87%. This model is really dangerous, the studious student who have written their writeup on their own will get negatively effected.

[LuvGyro]

now i can check if they know I'm cheating💀💀

[abibk]

Sir can't we copy past from chatGPT to Schoolsolver. what if i do so. Will they approve it or not. Is it legal to do so?🤔

[denzelsequeira]

Well i hope my professor does not get to see this video :)

[bdfact]

There is a simple non-AI solution: on very 1st day of class, ask students to write something on a paper. Make it like a game, turn the paper into four fold, ask them to write name on right corner, and then write in each square with subtitles like: who am I, expectation from course, things that don't like and things that I like. Collect paper and keep those for future reference. You may summarise findings of this exercise in the next class, without naming anyone. That handwriting narrative will show the level of language proficiency of the student. You can easily cross check home assignment with handwriting evidence. Another option is to do away with home assignment, just use photos to write stories in the class. Both creativity and originality will be easier to find from this exercise.

[leeishere]

I technically see chatgpt as a calculator but for essays.

[KT-_-]

when the teachers dont know we have access to the same zerogpt tools they do :)

[rodrigomgdealmeida]

GPTZero is not 100% accurate I already tested it and it sometimes claims it's AI and other times not. It's impossible to know if someone is using AI because in the final it's a written phrase.

[insaneinthemembrain]

if you have chatgpt write your entire essay and you do NOTHING to the text and simply copy and paste, you deserve the 0 lmao

[MementoMoriGrizzly]

Does this still work if the output is translated after?