@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.
English
Technical Communications Coordinator
732 posts
Technical Communications Coordinator
@TechCommMUNFEAS
Providing technical communications assistance to MUN's engineering students
Entrou em Ağustos 2021
115 Seguindo57 Seguidores
@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.
English
Welp. I've been expecting a scandal like this sooner rather than later
nytimes.com/2026/03/19/boo…
English
@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.
English
@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
English
@dowellml Just the in-text citations and not the reference list at the end, right? You're not manually creating reference list entries...
English
I’m still manually doing citations!
Shruti@heyshrutimishra
The new academic wealth gap isn't your university. It's not even your advisor's connections. It's who knows Claude can turn 50+ research papers into a thesis chapter in 3 hours and who's still manually coding qualitative data. I just watched a sociology PhD skip 8 weeks of analysis. Here are the 9 prompts they used:
English
@benjcartlidge @Nicole_Lee_Sch Both are issues. Ideally, peer reviewers would notice the low quality of AI writing. However, the desperation for peer reviewers has led to unqualified people doing peer reviews. They don't notice/care about AI writing.
English
@TechCommMUNFEAS @Nicole_Lee_Sch That's no the issue, the point is that AI written articles are being passed
English
If genAI can "write a paper for you" then at baseline you had nothing to say in the first place. LLMs cannot create new knowledge. They poorly synthesize whats fed into them. Good luck passing peer review with plagiarized work that adds nothing to the discipline.
English
@benjcartlidge @Nicole_Lee_Sch Exactly! People are waking up to the fact that unpaid labour in the form of reviewing articles is not worthwhile, especially when APCs can be thousands of dollars. It is harder to find qualified reviewers, leading to poor quality, AI-influenced reviews.
English
@Nicole_Lee_Sch But peer review is so broken that people are routinely getting through with work written by LLM
English
@Medievalhtybuff @jackcalifano It's funny. As AI-generated text becomes more prevalent, human writers begin to adopt patterns seen in the AI text. The result is that it's much more difficult to discern AI text from human text using AI tropes.
English
@jackcalifano I don't disagree but most AI text is pretty easy to spot
English
@Sally_Sharif1 I do feel that part of the problem is that people conflate "faster" with "better."
English
I disagree with the premise of this argument. AI CANNOT do social science research better than most professors. However, AI can easily plagiarize the research that has already been done by social science professors.
Alexander Kustov@akoustov
AI can already do social science research better than most professors with PhDs. And, for the first time in my life, I really have no idea what happens in five years. Things are changing already, we just need to wake up.
English
@Bayeuxpeen @MrEwanMorrison @UroborosJose In many ways, the oral exam is at least as important as the dissertation. The written compilation of their work is not the only way their work is presented. After all, many PhDs go on to lecture or teach in their areas of expertise.
English
@MrEwanMorrison @UroborosJose If the value of a PhD is knowledge production, and people can produce said knowledge in the form of a thesis with LLMs, why then should we value the oral exam afterwards? Sounds like gatekeeping.
English
This is not true. PhD students who write with AI are unable to defend their thesis in Viva presentations as they haven't worked through the processes of reading, learning, internalisation & argument formation.
Academics need to throw LLMs out, not 'wake up' & accept them.
Alexander Kustov@akoustov
AI can already do social science research better than most professors with PhDs. And, for the first time in my life, I really have no idea what happens in five years. Things are changing already, we just need to wake up.
English
@DrLancaster The entire system needs an overhaul. Reputable journals can't find qualified reviewers, so the road from submission to acceptance/rejection takes too long. The pressure to publish leads new academics to paper mills and predatory journals. It doesn't justify it, but I empathize.
English
The “pressure to publish” culture has led to suspect research papers being released faster than they can be retracted. I’m quoted in a #researchintegrity story which discusses the sheer scale of the research paper mill culture, an industry with so many parallels to student #contractcheating provision.
The link to the article is in the first reply.
English
@daniel_mcdowell In some interactions, I feel the person I'm speaking to having the urge to pick up a phone and ask AI to provide an answer to my question. Whether they know the answer or not, it is stifling communication.
English
I tell my students that one of the main reasons not to over-rely on AI for research/writing is b/c--at some point--you're going to have a face-to-face conversation with someone who matters, and if you don't actually know some things, that interaction will not go well for you.
English
@Marc__Watkins @aliciaandrz This should be printed and framed.
English
@aliciaandrz I think the issue is with AI there’s no longer a fear of the blank page. You can generate anything. But to write something it must matter and have purpose that’s meaningful to you. We’re going to have to advocate for this repeatedly going forward.
English
my husband is proofreading a book on writing & the author suggests using AI for writer’s block. he wants to add a note saying she should consider the implications of saying this—at the very least that she might be ostracized from the writing community. big 5 publisher. thoughts?
English
@mattgroh @NatMachIntell By measured, do you mean that certain soft skills can be assessed and given a numerical value of quality? I.e., can it be used in automated assessment with any sort of reliability?
English
🚨 New in @NatMachIntell 🚨 We collected 9000+ annotations of empathic communication in convos from experts, crowds & LLMs across 4 NLP/comms/psych frameworks
LLM judgment exceeds crowds' reliability & nearly matches experts
Soft skills can now be reliably measured by LLMs 🧵
English
@fake_journals I looked at one of these articles. It has an obscenely large font size and no in-text citations, yet a reference list is provided. Exactly half of the "references" are to his own works. Also, the "Version V1" in many titles is odd.
English
124 articles in 2025, no articles before that. Has an h-index of 24. Ethical?
Saw this researcher when we were looking at something else (we'll link to that post in the comments, once that post is live, but essentially, lots of self-citations).
This person publsihed (or at least indexed on Google Scholar) 124 papers in 2025. Before thet they had not published anything else (well, there are two undated articles)
Most of the papers are highly cited (self?), leading to an h-index of 24.
The identity has been blurred as some people do not think we should show it - but we have linked to the Google Scholar account below (you don't have to look), as it is public knowledge.
What do you think?
Google Scholar profile: buff.ly/mepKv66. We have archived this as an image as Google Scholar does not seem to allow archiving of it pages. The image is here: buff.ly/VnvFNd6
English
@Spottingthespot @SpringerNature It makes me wonder, in what training data are all these journal articles with one-paragraph subsections? 🤣
English
@DrLancaster A return? Did it go away?
English
With Wikipedia now likely to appear as a prominent source in AI generated content, does this mean a return to Wikipedia as a source in student work submitted for assessment? If so, another series of lessons on digital literacy beckon.
English
Do you contribute to Wikipedia, either as an editor or through donations? Your contributions are now being sold to AI companies, who Wikipedia is charging for access #academicintegrity
English
@fake_journals They missed the "L" in the URL.
ojs.albutana.edu.sd/index.php/jas
English
Always nice to be alerted to a citation to one of our articles, but some citations are better than others.
This one causes me some concern.
1️⃣ The way the citations are done is not correct. The authors cite using [n], yet they do not provide numbers in the reference list. As an example, our paper is [10] in the text but, but we have to count through the references to find the relevant paper.
2️⃣ I have tried to find the journal, but am struggling. I can't find it using search. The best I can do is: buff.ly/PetyiTw, which I extracted from the Google Scholar alert.
So, although I am appreciative of the citation, I wonder whether I should be worried about where I am being cited? They all count and this one means that our article has been cited 762 times.
===
Citing Article: Abubakr H.Ombabi, Mohamed Babiker Ali, Mussab E.A Hamza, Abuzer H. I
Ahmed,(2025), An Enhanced Metaheuristic Framework for Timetable Generation Using
Genetic Algorithm and Local Search , Al-Butana Journal of Applied Science (17): 31-53
Link to PDF (I got this ftom the Google Scholar alert): buff.ly/qMA0lll
Article being cited:Burke, E. K.; Kendall, G. and Soubeiga, E. (2003) A Tabu-Search Hyperheuristic for Timetabling and Rostering. In Journal of Heuristics, 9 (6): 451-470. buff.ly/U9oQC9E
English
@Marc__Watkins @emollick Absolutely! Using a calculator to determine an average will result in the same number if done by hand or by calculator. A pgh written by a human will be different from one written by either gen AI or another human. Writing is personal, so using AI to do it requires disclosure.
English
@emollick I think we first must establish the principle that it is necessary to disclose. It’s hard to do when people throw increasingly absurd suggestions that AI is like a calculator, so why would I disclose that? Normalizing disclosure should be the foundation of a socal contract re AI.
English
The lack of clear ways of disclosing AI usage is a problem for using them in many circumstances.
We need clear ways of acknowledging human contribution, from all human work to mixed work to directed AI to autonomous AI. Otherwise it is hard to assign credit (or blame) for work
Ethan Mollick@emollick
My modest proposal for how to acknowledge AI work is to use the Latin phase "Fieri Iussit," which means "commanded to be made," a common phrase on Roman Empire buildings. You didn't make the thing, but you commanded it be done, so the acknowledgement is "Ego hoc fieri iussi."
English
@Marc__Watkins Some blame should also fall on institutions that have AI-based tools as part of the software suite that the institution provides. There should be responsibility on the part of the institutions to provide guidance.
English
Many blame students for using AI to cheat, but we should direct our anger on an AI industry that markets itself as being safe and responsible while refusing to implement guardrails that stop students from using these tools to impersonate learning and committing academic fraud.
English
@Thatsregrettab1 @stevenemassey Agreed. The citation indicates that the information came from Grok. That said, the citation (source + URL) is not a citation style I am familiar with. The author may be confused as to how citations work. Any reader would assume that Grok is being used as a primary source.
English
@stevenemassey "Were found..." by Grok? Or by a search of the literature via PubMed or Dimensions, etc.?
If by Grok, how do we know that it didn't miss anything? Can that search be replicated? Will the results always be the same?
In my reading, the paper uses Grok as a cited source.
English
This preprint cites Grok as a source? Really?
I don’t know any serious journal that would publish this paper, and many/most editors wouldn’t even send it for peer review. No doubt some bottom feeder will take the author’s money though.
Steve Massey@stevenemassey
2/ A preprint describing the analysis has been 'on hold' for almost 3 weeks at arxiv, in another case of suppression of 'alert' reports. In the meantime, a copy can be found here: zenodo.org/records/179232…
English
@DrLancaster Some GenAI models heavily favour arXiv for academic sources. I'm assuming it's an access issue. I've definitely received literature reviews from students that use entirely arXiv sources.
English