Technical Communications Coordinator

@mayhewsw This is an area of debate at my institution. I have the same take as you: it can't be plagiarism because it is not copying the work of "another." For that reason, and because the process of detection is different, I prefer to call it "using an unauthorized aid" (i.e. cheating).

I’m going to regret getting into this, but would you call use of an AI tool to write a paper for you “plagiarism”? It’s definitely bad, don’t misunderstand me. But is it “copying someone else’s work”? I don’t think so

@GirliePsychosis @safiyaaaay It could be. If all the sources from one paper match all the sources from another paper in the same order, even if the wording is different throughout, it is likely plagiarized. Having one or some of the same sources as another paper is generally not plagiarism.

@safiyaaaay Citing the same works as someone else is never considered plagiarism. I actually don’t know what these fuckeroos are even talking abt. I’ve also seen straight up plagiarism that just slides (which is a much bigger problem than citation issue).

idk what the hell is going on in other disciplines but in History, lifting the citations from someone else's work is akin to plagiarism. You are falsely claiming to have done the research. It's intellectual dishonesty

@whitherapathy @eduleadership Certainly any cite-worthy paper would.

@eduleadership Exactly. If you cite it, you have to have read it. All papers contain enough information to cite them. There is no actual difficulty here.

This is extremely stupid because if you’re citing a paper, you should HAVE the paper and have READ the paper. You don’t need a database to cite a paper properly. You need the actual paper, and if you don’t have it and/or haven’t read it, you have no business citing it.

@mar_kar_ @pangramlabs Another point is that the reviews, unlike the article being reviewed, are typically not published anywhere for the world to see. There is no need for perfect grammar or spelling. If the feedback can be used to improve/reject the article, the goal is achieved.

Can we STOP generating REVIEWS? 1. 'I just polished' -no need to! also detectors like @pangramlabs would likely distinguish it 2. 'I just translated' - detectors actually do not pick up on translation + if you are reviewing EN paper you should be able to comment on it in simple EN. I miss imperfect human reviews and I'm a bit tired of 'but we can't be sure'... There are ways llms can help in review, writing review is NOT one of them.

@moorehn I really don't like classifying AI use as plagiarism in academic misconduct regulations. I'd rather keep them separate since there are differences in detection and, in some cases, intention.

An MIT writing professor on his students using AI: "I realized that for the first time as a writing professor, I had to deal with students producing words without work, which wasn’t quite plagiarism and wasn’t quite paying for someone else to do the job, but it felt like a kind of naive chicanery; a perversion of the contract between writer and reader." theguardian.com/us-news/ng-int…

@JeremyNguyenPhD AI overuses a common writing trope, people notice and either mimic or eschew that trope, and then AI moves on to a new one. AI writing is simultaneously causing some ppl to write like AI and others to eliminate certain previously acceptable phrases/styles.

I hope that LLM writing doesn't accelerate writing tastes and preferences to be like "fast fashion" in clothing. I can imagine an earlier time when clothes were all handmade and costly, and it was a bummer that lots of people simply didn't have enough clothing or the right kind. It's great now that we all have so much access to clothing, but it does seem a waste of human ingenuity and effort and frankly our lives for us to accelerate through waves of disposing of clothing when it stops signalling the right thing. I'll be a bit bummed if LLMs do that to our writing.

@emollick I do worry that its ubiquity will lead to humans adopting that style in their own writing, essentially training humans to write like AI rather than the other way around.

"Load bearing," "I keep coming back to," "Not X, but Y" A curse of using AI a lot is that you realize how much of the writing around you is just AI, now People who don't use AI have been unable to identify AI prose on sight, but those who use it a lot can spot the tells easily

@Marc__Watkins The problem with pen-and-paper writing assessments is that they move away from authentic assessment. No writing outside of school/uni is done in a timed session on paper. It is done in waves on a computer, using a variety of digital tools.

I was interviewed by Dana Goldstein for the NYTs about how writing instruction is responding to AI. It’s true many are returning to analog techniques to curb AI misuse. Some are also trying to incorporate AI to help students develop discernment about AI use.

@roman_janssen @fake_journals I've seen at least one instance of a DOI being reused by a journal: one article was retracted, and they gave the DOI to a new article. The IDF said that they can't really stop a publisher from doing it, but it is rare since DOIs are relatively inexpensive.

@fake_journals The cost of one DOI is about $2, but this also requires an expensive subscription to the DOI Foundation. While charging $500 for a DOI is fraud, even if you wanted to have a diamond access journal, you’d still have thousands in fix costs (like the overpriced DOI subscription).

How much does it cost to register a DOI? I received an email from the Journal of Neurology and Neurological Science (ISSN: 3068-7993). They said that there would be no APCs, but a DOI (Document Object Identifier) would be payable. I asked how much that charge would be. You can see what they said in the image. I note that it is "only" 500 USD and it is "towards" the DOI processing charge (suggesting that the DOI processing charge is more than USD 500). My questions are: 1) How much does it actually cost to register a DOI? 2) If you were going to only charge a "DOI fee" what would be a reasonable fee, given that it would be okay to charge a little extra to cover your own administrative costs? Journal URL: jnans.org (archived at buff.ly/EaJuulv)

@NC_Renic I've had to reduce the marks I assign for grammatical accuracy and other language components. AI has eliminated many of those errors. Then you get the student who struggles with language, and you have mixed feelings about the mark they get for a hard-to-read paper w/ good ideas.

AI cheating has gotten so bad that I now feel genuine affection for horrifically bad essays clearly written by the student

@rwlesq @keysmashbandit I do feel that the style of the one on the right is better. The one on the left refers to three things as "basically ____." The ChatGPT version is more formal (if that was the intent). The hedging is part of the formality, but I agree that it is overhedged.

@keysmashbandit There's a reason why I didn't just cut and paste ChatGPT's response, but I think the final product is better having run it by ChatGPT and considered its input.

Please, I'm begging you, try to critically examine the differences between these two pieces of writing. ChatGPT editing did not improve this. Every single change only served to weaken your claims significantly. Everything is now hedged into oblivion: no longer have you outlined a "problem," now it's merely a "flaw." "It is true" now demoted to "it appears to be the case." "Is" gets a "usually" tacked on. A thesis statement at the end of the first paragraph gets run over by noisy, out-of-context example-whittling. All for fear of being misconstrued. And at the end, the argument that gets spat out isn't even yours anymore! You argued that Graeber failed to create a true account of work because he did not understand Chesterton's Fence. ChatGPT is arguing is that it is possible some apparently bullshit jobs could be secretly load-bearing if you squint. These are two different statements. The second is weaker and less compelling. It says less. And it's fucking longer! Don't do this anymore! Stop doing this! It's worse!!!

@dowellml I do wish you had polished it, though. I find it hilarious how the "polishing" so often results in unnecessary bullets and emoji.

No, Google AI, I do not want to “polish” this email which is a list of free agent bids for a fantasy baseball league. There is no polish present, needed, or available. It is not worthy of polish nor would it benefit from polish. What the fuck do you even want to polish?

@MrEwanMorrison Bad: using AI to write an article, creating fake references in the process Also bad: citing a source you found in an article, and that source turns out to be hallucinated The difference: the second one often doesn't know what they are doing is wrong

110,000 papers contained AI invented references in 2025. This is the pollution of science by AI slop.

@NatalieO73314 I think, as with most tools, there is a danger of the user just not knowing how to use it properly. The Correctness tab, as indicated by everything underlined in red, would only fix language errors. By default, Grammarly suggests many revisions, some of which are not "errors."

@TechCommMUNFEAS If I’m being generous, perhaps its functionality has improved since this happened. It was making very basic mistakes at the time. But one usage I’m very strongly against is using it to ‘check’ a professional editor’s work!

A big publishing client refused to sign off on an edit they had commissioned me for until they had run my work through Grammarly. Many of the issues Grammarly ‘fixed’ were incorrect. It was one of the most depressing — and, frankly, demeaning — moments of my career. 🥲

@Marc__Watkins @TheLincoln I concur with your assessment, @Marc__Watkins. Either way, AI was misused, but this similarity would probably require asking for a rewrite of the existing review rather than a new review from the ground up. I also think this being an AI rewrite of an original review is unlikely.

@TheLincoln I think a much more likely explanation is the reviewer took the Guardian article loaded into the context window and prompted it to write a new review based on the original article as a seed text. That’s much more likely.

For years, LLM evangelists have claimed they are incapable of plagiarizing. But what else would you call this?

@MereSophistry When I ask AI to write academically, it consistently has a longer average word length (typically 1–1.5 letters longer) and usually a longer average sentence length (25+ words per sentence, on average). "Make it longer" does sound like a goal a confused student might have.

There is a massive problem with bad writing. It’s in academia, management, Silicon Valley, entrepreneurialism, and other domains. AI accelerates this problem because it writes in the way a first year college student imagines they should write in academic jargon.

@ednewtonrex @FinanceDirCFO @MrEwanMorrison The problem, though, is that no proof is provided. With traditional plagiarism detection, it willshow you the article with similar wording. With AI detection, it's basically just saying "trust me."

@FinanceDirCFO @MrEwanMorrison Some actually seem pretty good, e.g. Pangram (which gets this one right). In general AI text detection is very useful IMO, particularly as many people find it hard to detect themselves. May not work forever, but right now it’s quite good at identifying slop

AI detectors can't even detect the presence of AI properly. If they can't even do that, I'm not inclined to think they'll be doing anything useful any time soon. The entire sector is a con...

@Marc__Watkins @TheLincoln @mavisclare I would imagine that book publishers will eventually require document tracking. Some jobs currently track employees' activity when working remotely, and this should be treated in the same way.

@TheLincoln @mavisclare This isn’t going to be sustainable for publishers to pull books they suspect are AI generated. At some point we have to talk about where the burden of proof lies in making these judgements. I’m all for disclosing if you use AI, but if that leads to being cancelled few will do it.

Welp. I've been expecting a scandal like this sooner rather than later nytimes.com/2026/03/19/boo…

@kylejunlong @majamediaco It is definitely happening. It is worth studying, in my opinion. I know that I opted for commas nearly 100% of the time before AI. I now use em dashes every now and then, but certainly not to the extent that they are used by AI.

@majamediaco 100% reads like AI, although i wonder how much people’s writing has been shaped by LLM usage such that someone who didn’t write much / have a distinct style before AI just writes like that because it’s the type of writing they engage the most with

found a blog with 16,000 subscribers where every post feels AI-written then found a post adamantly denying they use AI… which also reads like AI to me what is happening