What to do if you are falsely accused of cheating with AI

I have started seeing posts nearly every day on Reddit from students who say they are being falsely accused of using ChatGPT or other AI tools on their assessments. I have also had several students reach out to me directly after seeing my video about false positives to get more information. In this video I run through three important steps to take before a meeting or hearing.
My blog post on where I will collect additional resources and research
drlyndonwalker.com/what-to-do...
Do I Write Like a Robot: False positives with AI Cheat Check video • False positives from a...
My SkillShare courses: (includes a 1 month free trial)
Getting started with R skl.sh/3cyKwzi
How to Write an Academic Journal Article skl.sh/3hpm3i0
How to read an academic paper skl.sh/3wO9Skp
Introduction to Flipped Classroom Teaching skl.sh/3TBK7jw
Connect with me:
🐥 My Twitter - / drlyndonwalker
💻 My Website - www.drlyndonwalker.com
Join my Facebook Community - / drlyndonwalker

[@MadamKorgor]

How ironic is it that humans are using AI to detect AI writing to accuse humans of using AI generators to cheat?

[@SnagKing25]

You can’t make this shit up

[@JimmySonny]

I find these AI detecting tools unethical, unless it is to detect plagiarism of another's human work. If you want to know whether someone knows what they're talking about, have them explain it in person, on a call, and see if they can argue and elaborate on their work. Some may use AI to do something they can already do but would take longer to do themselves, like coding, and I don't see why we should be punishing people for being more efficient.

[@DrLyndonWalker]

I don't think the detection tools are inherently unethical but the way that they are applied and the way they are marketed definitely can be. Having someone explain in person or on a call is a step you can take but this does disadvantage non-native speakers and those of us who are neurodiverse. I prefer to set work where you can't "cheat" as such because you need to show the process of the work, not just the final outcome. That could also include activities like using tools like ChatGPT and then using your brain/skills (eg. get it to write some text or do some coding then you as the future expert have to evaluate it). Thanks for the comment :)

[@JimmySonny]

@@DrLyndonWalker Yeah, sorry, the use of the tool can be unethical, you're right. I agree with the rest of the comment, especially the last paragraph.

[@HelloOnepiece]

@@DrLyndonWalkerI would also add that quite a low number of students actually remember what they wrote in their papers after 6 months, or at least not much more than AI

[@rikiriki43]

Thank you so much. As an international student. I was accused AI. However, I didn't use it. It's very stressful for me to be accused. I am afraid to have a hearing with professors, it's horrible.

[@DrLyndonWalker]

I'm sorry to hear that. I hope my resources are helpful, feel free to share them. The article I link on my website about non-native English speakers will also be relevant to you.

[@alexmason2595]

Hey thanks man this is really helps, I’m glad there is some hope. It’s really upsetting when you work hard for something just to be denied over something you have no relation to. I will be sure to keep these point in mind.

[@sebastiancastro3253]

Thank you so much, my college is currently making me go through a formal hearing to defend my case as I've been charged and/ or sanctioned for "AI plagiarism". It's incredibly frustrating especially since I have no previous version history of my document and don't want to lose my scholarship. I use grammarly, and I suspect that that is what made my teacher suspect me. Wish me luck!

[@DrLyndonWalker]

I'm sorry that this has happened to you. I have had another student reach out who might lose their military scholarship if found guilty. It definitely looks like Grammarly increases the chances of a false positive. I hope my video and website content can help in your defence. Good luck.

[@pepe_s1lvia]

Hi, I'm hope I'm not too late. I'm not sure what detector they used, but they sure used one. I'd suggest you getting copies of work or papers that your professors have published, and tell them to run it through the detection software without getting flagged for AI. I feel for you. This is totally unfair, since AI is doing more harm to the people that don't use it than to the people that do use it. There are currently a few AI detectors that I know of, which are ZeroGpt, and Turnitin's own BS "98% accurate" AI detector. Even the Constitution gets flagged by AI detectors. My College accused me as well, but were much more lenient, since another student was detected for AI, at about 4%. He amended it on the spot in front of the professor, and it got marked up to 20% AI. The fact that they chose to charge you and go through with in on mere accusations shows the absolute BS wankery of education system. They chose to fear AI, knowing that this would change everything, instead of embracing it. Contact me further if you need any more ideas or have any questions. Publicly shame your teacher or charge them as well, since it is an act of defamation, which threatened to ruin your life. Don't worry about the legal fees, because if you don't fight back and let them win, you'd be wondering how you're ever gonna have a future. Good luck

[@CyberDancerDany]

If I were you show your web search history when you did your assignment, that can prove AI wasn’t used. Grammarly is to help you make better sentences, it is not something that writes an essay for you like ChatGPT does. It’s no different than getting your paper proofread by the professor or at the schools tutoring center.

[@6moonsofsaturn]

@@CyberDancerDany Professors and tutors shouldn't be proofreading and editing papers for students. That is different than providing feedback on higher-order writing concerns such as structure, adherence to the prompt, thesis development, analysis, demonstration of critical thinking, etc. Having someone (or something) completely rephrase and rewrite student work IS cheating, and the final product is then not a true reflection of the student's writing. No reputable writing center would allow a tutor to edit work for a student. Tutoring and editing are not the same thing. That is why Grammarly is problematic and why in-class writing will likely be making a huge comeback, sadly.

[@Sleazze]

Hey Sebastian, what ended up happening? Did you prove it wasn't AI?

[@sowathanakhamphean8356]

It happened to me! I don't know why or what happened!? I failed the whole course.

[@ydherdn]

Thank You

[@lokojo3184]

Hi, I was accused and it has been a stressful experience. I appreciate this video. I wish AI didn't exist because this has taken up a lot of the time I should be working.

[@DrLyndonWalker]

I'm sorry that happened to you. I hope the word gets out further and faster on the issue of false positives as I'm seeing far too many instances of false accusations.

[@cowboydoggo8200]

I got accused today and a majority of my essays were given an F. I have screen recordings of my most recent one essay along with questions I had prepared myself to make my essay great. (I delete older ones because I have no memory/space on my device) I emailed him saying I have proof that I did not use A.I for and if he's willing to look at the recordings. Now I'm anxiously waiting for his reply and what happens next. 🙃

[@carultch]

The loophole is to have AI generate your work, print it out, and retype it while doing your screen recording. There is no such thing as proof you didn't use AI, other than proof that it was written before AI was invented.

[@Jennifermommy]

i used ai only for research and that research was wrong so my essay is flagged ai but i did lie that i used ai because i was scared now i am going through a meeting and i am scared

[@user-jn9pg8gs1p]

ur the best

[@JacobWindle]

I am an AI researcher currently, working in the education space. I wonder if we're going about this all wrong in terms of AI detection. Current research has shown that AI detection is about as good as a coin flip, especially with generated text. And, students are learning ways to prompt the AI to sound more like them, given a few examples of their own writing as part of the prompt. What if AI detection worked more to augment a student's understanding? For example, if the system thinks that AI-generated text was used, or maybe a student pastes an answer into a text box, could we ask for clarification of an answer? Maybe alert the teacher that AI is being used to help understand if a given student might need more help? Another commenter pointed out explaining things in class to help with cheating. I've heard of that being used quite a lot by local instructors in my area. Just wondering what additional help we could add to make teachers lives a bit easier and lessen the stress on students.

[@DrLyndonWalker]

You are incorrect with your "coin flip" accuracy statement. Some are up to 99% accurate but for a large pool of students this still creates problems with too many false positives. I don't think AI detection will ever be the answer. Instructors should be designing assessments that demonstrate proof of work (eg. drafts, recording, process etc.) and tasks that are hard to be done well by an AI (or better yet, also including AI use then demonstrating the need for humans in the process)

[@patjohnsonofficial]

I find it amusing how they can formulate seemingly accurate statistics regarding false positives, yet they can't ever provide evidence of using AI generative programmes to begin with 😂

[@JoeBurner1720]

My professor thinks using Grammarly as a spell checker counts as AI…

[@oldmedstudent1750]

It literally is AI. Using AI doesn't only mean creating the content, it also means editing your content on your behalf.

[@JoeBurner1720]

@@oldmedstudent1750 I know, but he doesnt know the difference

[@newkiller4457]

@@oldmedstudent1750thats ridiculous since majority of colleges allow grammarly as it helps you fill out those typos.

[@oldmedstudent1750]

@@newkiller4457 grammarly literally requires your papers