Stop using AI to submit your work
I am putting this information out to all my classes, both face-to-face classes and online classes.
It is very clear to all of us that some students in this class are using AI to complete assignments, including the assignments and discussion board questions and responses. By all of us, I mean me, other professors, and students. Students have contacted me about other students using AI in discussion boards. How do the students know? Because the AI submissions are crap.
I routinely use ten(10) different AI detectors to check for this content once I see what I suspect is AI-generated text.
What are the consequences of using AI to submit your work
Using AI without attribution, explanation, or permission for assignments is cheating. It is plagiarism and collusion. Read CTC’s Academic Misconduct Policy for further clarification and explanation. https://www.ctcd.edu/academics/catalogs/catalog-texas/academic-policies/academic-misconduct-policy/
The consequences of using AI to submit your work can start with receiving a zero on the assignment and can lead to getting dropped from the class.
What is going to happen to me if I have already used AI in this class?
If you have already used AI, I may have already graded your assignment and may have made some comments about your language being too broad, your failing to number and specify your answers to particular questions, and non-specific or something like that. Those are clues. Those clues and this announcement mean I am offering you an off-ramp and providing you with a solution and an opportunity to do your work differently and/or better. Even an opportunity for you to learn how to use AI properly. I am still evaluating what I am going to do with violations of CTCs policy for these submissions and doing more examination of the submissions.
Here are some things you can do that will help your case.
Message me and ask me if you can submit an assignment again – no questions asked. I will cancel your attempt, remove your current grade for that assignment, and allow you to resubmit – no questions asked.
Tell me you used AI, what you learned from it, and that you want to resubmit or modify your submissions. No judgment at all from me. AI is new and students sometimes don’t know what to make of it.
Do nothing and continue to use AI without attribution and notification. That can result in negative consequences for the student.
What does AI generated text look like?
AI looks like beautiful crap. AI does NOT do a good job of answering the questions being posed to you in this class. AI uses broad, sweeping language that is grammatically correct and even a bit poetic but horrible at answering a direct question. The bottom line is that each of you (students) can spot this in the responses you are getting from other students and some of you already have spotted it and asked me about it. Check out some of the feedback you are getting from other students. It doesn’t look like a human wrote it and it probably doesn’t even do a good job of giving you feedback. Take what I am writing to you right now, this sentence, and you can tell that I am a human. I am using complete sentences and mostly correct grammar, but not perfect. I am directly addressing you and directly addressing the issue at hand. The sentence I began this paragraph with, AI looks like beautiful crap, is the most succinct way I thought of to describe to you what AI appears to be from your professor’s point of view. When I ask students for answers and feedback, I don’t want abstract art. I want direct answers in your own words.
AI causes students to doubt themselves
In some cases, AI feedback may make some students feel inadequate because the vocabulary appears to be rather advanced and creative causing students to feel like they are out of their league. Don’t feel that way. If you are using your own words and doing your own work you will be fine.
Using AI to do good work is difficult
Students sometimes operate under the delusion that they are the only people on the planet with access to the internet. This is apparent from their efforts to cheat and take shortcuts. Usually, the effort is sloppy and poorly executed. What I mean by that is that students don’t do a good job cheating. AI is a powerful tool and will benefit all of us, and I encourage students to use and study it. However, using AI for work and school requires a lot of effort and editing, and it is very clear from the students who have already tried to pass AI by me that they have no idea how to use AI.
I conducted research and wrote a paper called “Detecting AI” this year. I will provide you guys access to that paper, and you will likely learn more about AI and how it works. It is only a rough draft and as with AI, it must be modified constantly.
During that research I did for that paper, I used the paid and free versions of AI Chat bots including Chat GPT developed by Open AI. I also tested these AI checkers thoroughly and researched exactly how AI works and why it is horrible at performing certain task without a lot of help.
Why students are NOT very good at using AI
If you have used and tested AI as much as I have, you realize that it is sometimes difficult for you to get AI to actually understand and answer the question you are asking. I am very clear and have a deep understanding of procedural law. So, I know what exactly to ask AI and I also immediately know whether the answer AI provides me is good, total crap, or somewhere in between. Even if AI gives me the correct answer, it may not exactly answer what I asked specifically. In other words, it got me in the ballpark or generated a foundation, that I must now edit and modify.
Here are reasons why students are so bad at using AI
Students don’t understand or take the time to read or understand the question being asked or the material. Students blindly paste text into Chat GPT and don’t even understand what they are asking Chat GPT. I have spent several minutes just crafting the correct wording for what I use to paste into Chat GPT. In other words, if you put crap in, you get crap out.
Students don’t understand the material and therefore don’t know if Chat GPT has answered their question correctly or if it has even answered their question at all. Sometimes I have had to ask Chat GPT five follow-up questions to narrow its focus enough so that it answers my original question. It is obvious that students don’t do that.
Chat GPT can’t properly source information it provides you. That is one of the things I noticed right away when experimenting with Chat GPT. I had to ask several times for sources, and it did provide some, but they were incorrect. To understand why that is, you must understand how language models and AI work. I will briefly address that in my paper.
Using AI properly is hard work. Cheating, plagiarizing, and colluding is hard work if you want to get away with it. In order for students to use AI well, they must know the material well enough to know if they receive the correct answer from Chat GPT. If students know the material that well, they don’t need Chat GPT.
An old professor of mine used to allow us to use cheat notes in class on any exam, but he required us to show him our cheat notes and restricted us to using a single 3x5 card (front and back) and nothing more. It required us to write very small and be selective with the words we used for economy of space. What we did not understand at first was that making this cheat card actually required us to understand the material well enough to summarize what was relevant. The note card just helped us learn and we ended up writing more on the exam than what was written on our card because we knew the material when we finally took the exam.
CTC Professors share information with each other
In the first week of class, I have already had conversations with professors from other departments at CTC in which we shared with each other some of the ridiculous submissions from students which were obviously the product of AI. Some students even used AI to write a paragraph about themselves, and some students forgot to remove the AI prompts generated by AI.
Examples from student already this semester
For example, one student submitted a sentence that began like this, “My name is [your name here] and I am excited …”. In some of my classes this semester, students have submitted paragraphs of information and none of the words or sentences came close to what the question was asking. The question was about procedural law and Chat GPT was talking about procedural law, but nothing in the answer came close to answering the question. The information Chat GPT returned was correct, but it was talking about apples and the question asked was asking about oranges.
What we want from students at CTC
First and foremost we want you to become better, smarter, and to prepare you for the real world. So my first concern is verifying that what students submit is actually their work. I am not looking for something beautiful to blow me away. I am looking for your honest output and effort so that I can provide you with feedback and help you get better at whatever you are going to do.
Students who want to use Chat GPT
I am not opposed to students using Chat GPT. If you do, you must declare that you are and use correct APA style. I am including some links on how to do that below. In practical and simple terms, put quotes around the AI language you use, and then critique (in your own words) what AI said. Tell me it got it wrong or right or if you disagree with it or if it completely misunderstood what you are trying to ask.
AI can help you learn
AI can help you learn. AI can help summarize things. AI can explain concepts in an almost infinite number of ways instantly. I will even provide some examples in this class of questions I pose to AI, how it answered, and then evaluate it. I can’t stop you from using it and therefore I am telling you how to use it correctly. I am also telling you that the way students are using it now is not working and it can have negative consequences for the student.
So do NOT use AI as your own work. That is cheating. You can probably get away with it for a bit, but it won’t work forever and everyone is aware of it – even your fellow students.
How to cite ChatGPT - https://apastyle.apa.org/blog/how-to-cite-chatgpt
How to Cite ChatGPT and AI in APA Format - https://www.grammarly.com/blog/ai-citations-apa/
Citing Generative AI - https://guides.libraries.psu.edu/apaquickguide/generativeAI
APA Citation Guide (APA 7th Edition): Artificial Intelligence - https://library.senecacollege.ca/apa/artificialintelligence
Let me know if you have any questions, and remember that we are all here to learn and it is my job to help you.
Thanks!
Scott Lorenz