A whole new ballgame in cheating. Introducing ChatGPT

[Wahoo Redux]

And:

Fortune: A Wharton professor gave A.I. tools 30 minutes to work on a business project. The results were 'superhuman'

[Diogenes]

I had a reference list that was a mess to clean up so out of curiosity I had Chat GPT fix it like a citation generator/spell checker. The results were meh. It helped a bit, but I still had to go in and fix every reference somehow. So it saved me a small amount of time and tedium. This is how I would let my students use it.

[Sun_Worshiper]

I've been using the Bing AI chatbot quite a bit lately and it has definitely given me a productivity boost. It does so in two ways: (1) It can write emails, article summaries, or even short essays. These then need to then be edited/retrofitted for purpose, but it still provides a nice template and some filler text that can save me quite a bit of time. (2) It works as a glorified search engine. For example, I could ask it to find me statements from policymakers on the legality (or lack thereof) of drone strikes or strategies that companies take to deal with bad press along with examples. From there, I can ask it follow up questions to get a more precise answer or additional details. This has to be fact checked, since the info it gives is often wrong (especially if I put it on "creative" mode), but it still saves me a ton of time compared to trying different search terms in google and then having to skim through documents.

I also have some ideas about how I might use it for grading, although I haven't had an opportunity to test it out just yet.

[Sun_Worshiper]

Anyone using ChatPDF? I definitely recommend playing around with it if you haven't already.

[Liquidambar]

Interesting idea.  I dumped in the most recent paper my research group read, but I wasn't that impressed with how ChatPDF did.  It correctly answered a couple questions about details but kind of missed the main point.

[Sun_Worshiper]

Interesting idea.  I dumped in the most recent paper my research group read, but I wasn't that impressed with how ChatPDF did.  It correctly answered a couple questions about details but kind of missed the main point.

I used it to identify a few questions for students on an article I assigned them, which I haven't read in several years. It did the trick quite well. But ymmv depending on what you are hoping to get from the document.

[apl68]

ChatGPT is making it hard for paid cheaters academic writing assistants in Third-World countries:

For the past nine years, Collins, a 27-year-old freelance writer, has been making money by writing assignments for students in the U.S. — over 8,500 miles away from Nanyuki in central Kenya, where he lives. He is part of the "contract cheating" industry, known locally as simply "academic writing." Collins writes college essays on topics including psychology, sociology, and economics. Occasionally, he is even granted direct access to college portals, allowing him to submit tests and assignments, participate in group discussions, and talk to professors using students' identities. In 2022, he made between $900 and $1,200 a month from this work.

Lately, however, his earnings have dropped to $500–$800 a month. Collins links this to the meteoric rise of ChatGPT and other generative artificial intelligence tools.

https://restofworld.org/2023/chatgpt-taking-kenya-ghostwriters-jobs/


As has been discussed above, ChatGPT can't produce truly good-quality academic writing, but it can grind out mediocre as-long-as-it-passes work for free, so there's no longer any need to hire an actual person to do it.  Even at Third-World wages.

[mbelvadi]

From that article: "Collins now fears that the rise of AI could significantly reduce students' reliance on freelancers like him in the long term, affecting their income. Meanwhile, he depends on ChatGPT to generate the content he used to outsource to other freelance writers."

Oh, the irony!

[history_grrrl]

ChatGPT is on my shitlist.

I have a student in two classes this term.  He submitted a paper proposal that came in at 100% AI, but we had no policy and there was no way to prove it; it was grammatically perfect, though, and while it made correct though very broad, generic claims, there was no indication that Stu knew anything concrete about the topic though we had covered it in class and had readings about it. Also, his sources were unrelated to his topic. So it got a D.

Then I read his final paper and realized his quotations were made up, his footnotes didn't match anything he was citing, and one of his sources was nonexistent. Met with him, and it was an absurd charade of him acting all puzzled about where those quotations came from ("Oh, yeah, that was from a website I found with a lot of documents on it, I think").

Then I read his final paper for the other class, and surprise! Every quotation, every citation - made up. This was a paper based on letters people wrote long ago. AI makes up quotations from letters that don't exist. Names letter writers who don't exist.

And then I read both final exams. These are trickier, because they contain almost no detail. But when they do, they're wrong. "As discussed in seminar" - a topic never discussed. "In X film, Important People Y and Z share their reflections on the era" - nope, one appears in archival footage, and the other consulted on the film, but neither is among the talking heads.

I am disgusted by this student, who thinks he'll be going into teacher ed placement in the fall. But I have to admit, I am glad to have had my eyes opened. I think I'm one of those people who thought AI was sort of like a giant version of Wikipedia. Nope. Just ask it to write a biography of yourself.

[AmLitHist]

^^ This. ^^ So far I've had one, and the student caved immediately.  The "tell" was the illiterate word salad BS of all her work to date, only to have the most recent paper pretty much flawless in its writing, though vague and ephemeral in content.  Nope.  Not going to fly. 

I didn't yet have any specific anti-AI language in the syllabus, but I always include the line, "It should go without saying:  any work you submit with your name on it is presumed to be 100% your own work.  If it is not, the submission will be treated according to the course plagiarism policy" (which is a zero and academic integrity violation report on the student's permanent file; second offense gets an F for the class along with the report and a request that the VPAA move to suspend the student).

It's not like we haven't gone over the academic integrity rules roughly a dozen times to date.  I don't feel bad at all for busting a plagiarist, for AI or otherwise.

[Hibush]

^^ This. ^^ So far I've had one, and the student caved immediately.  The "tell" was the illiterate word salad BS of all her work to date, only to have the most recent paper pretty much flawless in its writing, though vague and ephemeral in content.  Nope.  Not going to fly. 

One suggestion from our admin is to allow students to use ChatGPT to create a first draft, then edit that for the content that ChatGPT is incapable of. The purpose is to skip the illiterate word salad stage so that the student can engage more with the learning. Instructors seem receptive at least to not have to read so much crud. The expectation for sharp arguments may rise.

[Parasaurolophus]

That's the suggestion here too. As if the word saladers aren't going to have the AI do their editing for them. Sigh.

Next time I teach ethics, though, I'm going to demand that my students plagiarize their first essays. They'll have to answer the question correctly (and it will be one that the AIs struggle with), and they'll have to be able to prove it was plagiarized. With a goodly number of points for ingenuity and successful masking of the plagiarism, as well as for a discussion of precisely what the harms are, and to whom. My hope is that they'll be sufficiently daunted by the difficulty of doing an adequate job that it will jar them out of the habit.

[phi-rabbit]

I teach four different classes right now (most intro-level gen-ed classes) and three of them are new preps.  I can't take the time to do all the things our University's teaching center thinks will deter plagiarism, if they will even work at all.  There's only so many ways to craft an introductory level paper prompt in my field and there's only so much grading I can do.  I absolutely despise cheating and punish it harshly, when I can be sure enough about it.  I also usually spend a disgusting amount of time at the end of each term figuring out where various plagiarized papers came from and documenting them.  But this semester I think I'm going to have to be done trying so hard.  That's the other advice our teaching center gives: stop trying to hard to catch cheating.  It kills me, but OK.  I'll give lousy grades to the lousy papers they're going to come up with and move on.

I already eschew papers and give only in-class essay exams when I can.  Unfortunately my university's gen-ed program requires some classes to have a certain amount of formal writing and so I cannot completely avoid assigning papers.  If I could, I would.  I do not believe there is any real value in assigning papers in intro-level classes like I mostly teach, in part because of the high incidence of cheating.  It wastes my time and provides no benefit to the students.

[apl68]

That's the suggestion here too. As if the word saladers aren't going to have the AI do their editing for them. Sigh.

Next time I teach ethics, though, I'm going to demand that my students plagiarize their first essays. They'll have to answer the question correctly (and it will be one that the AIs struggle with), and they'll have to be able to prove it was plagiarized. With a goodly number of points for ingenuity and successful masking of the plagiarism, as well as for a discussion of precisely what the harms are, and to whom. My hope is that they'll be sufficiently daunted by the difficulty of doing an adequate job that it will jar them out of the habit.

Good for you for trying to turn the whole business into a teachable moment.  And good luck!

[Parasaurolophus]

That's the suggestion here too. As if the word saladers aren't going to have the AI do their editing for them. Sigh.

Next time I teach ethics, though, I'm going to demand that my students plagiarize their first essays. They'll have to answer the question correctly (and it will be one that the AIs struggle with), and they'll have to be able to prove it was plagiarized. With a goodly number of points for ingenuity and successful masking of the plagiarism, as well as for a discussion of precisely what the harms are, and to whom. My hope is that they'll be sufficiently daunted by the difficulty of doing an adequate job that it will jar them out of the habit.

Good for you for trying to turn the whole business into a teachable moment.  And good luck!

It's going to be a miserable failure. But I hope it'll be more fun to read!