“university leaders … must act to help us collectively turn back the tide of garbage software, which fuels harmful tropes (e.g. so-called lazy students) and false frames (e.g. so-called efficiency or inevitability) to obtain market penetration and increase technological dependency” 3/🧵

🧵 Thread with my favourite quotes from: 📝 Guest, O., Suarez, M., Müller, B., et al. (2025). Against the Uncritical Adoption of 'AI' Technologies in Academia. Zenodo. doi.org/10.5281/zeno... 1/🧵

"Ultimately, these systems cannot really replace humans, replace the quality of human craft and thinking — so many of their capacities are overblown and displacement will only happen if we accept the premises (@olivia.science, 2025). .” zenodo.org/records/1706... 38/🧵

I find that my students really want this - they want to do things as humans. It flies in the face of what we are being handed down by the administration. But I will keep pushing to show Ss they are being robbed of their learning if they hand it over to something else...

Since someone else posted my ai statement from my syllabus on the thread, sharing myself as well. Need to push back… bsky.app/profile/nsou...

Gonna showcase your & @olivia.science latest paper to my ~324 stats students next Tuesday. 🙏 (While explaining the university's policy on 'AI' usage...)

Same

As a professor myself, I’m disgusted by any prof using so-called “AI” in the classroom.

I LOVE this one.

I realized years ago that Oxford and Cambridge had sold out to Big Tech when Bostrom's Musk-funded Future of Humanity Institute popped up in 2005.

For clear visual reference, the study and some great highlights from it here: bsky.app/profile/oliv...

Getting as many admin and other higher ed folks to read this and implement some of its suggestions as possible is a vital step: bsky.app/profile/oliv... Re: students, I think zero-tolerance policies are the way to go, like any other cheating. Sucks enforcing without admin support though.

If you, like me, are the parent of a neuro-atypical child dreading this autism press conference, please remember: These are the people hurting our kids. Not us. They are the ones stigmatizing and pathologizing them. They are the ones peddling untested science. And we should be very, very angry.

My child, a sophomore English major, tells me her classmates are appalled at how much their professors are promoting and using AI, themselves.

Yup and the Barnum effect too en.m.wikipedia.org/wiki/Barnum_... as well as how these systems are literally designed to make you think they are reasoning that's literally their only usp

Lesenswert in dem Zusammenhang natürlich auch der Artikel von @irisvanrooij.bsky.social zum Einsatz von LLM-Zusammenfassungen auf Websites von wissenschaftlichen Journals: irisvanrooijcogsci.com/2025/08/12/a...

www.rosa.gr/opinions/to-... Έγραψα μερικές κουβέντες για την πολύ προβληματική ανακοίνωση συνεργασίας της ελληνικής κυβέρνησης με την OpenAI. / I wrote an op-ed on why the announcement of the collab btn the Greek govt & OpenAI for ChatGPT in our classrooms is *BAD*

If you're observing this, I think you aren't being critical in your follow-up questioning. A useful lens is that there is a plagiarism-to-bullshit slider. If nearby responses were in the training data, you get more plagiarism. If not (or if "temperature" is higher), you get more bullshit.

This is the cognitive bias. It's related to what Weizenbaum observed with ELIZA. If you ask follow-up questions of the sort you might ask a student, it'll fall over hard. (You've probably seen endless examples that a preschooler would ridicule, e.g., kieranhealy.org/blog/archive...)

The abstract opens with flagrant nonsense. An LLM has no representation of ideas, certainty, or factuality. "Hallucination" is not an aberration, but inherent in every response. The magic trick is that remixing words somewhat out of context and without understanding can often be perceived as true.

Why is the output sometimes perceived as useful despite not being the product of ideas? Perhaps someone wrote similarly on a related topic and it "fits" to mix the words. A magic-8-ball or ouija board could nudge toward understanding, yet the harms of misplaced trust abound. bsky.app/profile/iris...

Yet again, machine learning — even gussied up via the transformer architecture — encodes and reinforces societal biases. This study reveals that LLM-based peer review relies heavily on author institution in its decisions. arxiv.org/abs/2509.15122

I want technology to do things and solve problems, not do a worse job at the parts of humanity that are the most human, like writing and art.

I understand AI isn't going to disappear and I work with information professionals so it would be foolish for my colleagues to ignore it, but when my colleagues casually use it for a creative endeavor (especially producing images) I feel ill.

Getting as many admin and other higher ed folks to read this and implement some of its suggestions as possible is one possible step! bsky.app/profile/oliv... Re: students I think zero-tolerance policies are the way to go; bullshit to have to try to enforce that without admin support though.

quelle surprise that academia dot edu now offers AI slopcasts they were a garbage flattery service ten years ago, and they have turned up the enshittification a few notches ideophone.org/academia-edu...

quelle surprise that the same shitty company is now taking all rights to turn your academic content and likeness into AI sludge in your name, unless you opt out

“HONESTY implies that we do not secretly use AI technologies without disclosure, and that one does not make unfounded claims about the presumed capabilities of AI technologies.” zenodo.org/records/1706... 27/🧵

“RESPONSIBILITY precludes scientists from using AI products whose use is irresponsible, e.g. harmful to people, animals, and the environment, or otherwise in violation of legal guidelines (e.g. copyright, data privacy, labour laws …)” zenodo.org/records/1706... 31/🧵

“INDEPENDENCE means that scientists ensure that their research is unbiased by AI companies’ agendas, and that any potential conflicts of interest are declared in publications and other public communications” zenodo.org/records/1706... 30/🧵

Full paper here: zenodo.org/records/1706...

A magic eight ball

OpenAI laughs in "open" does-not-mean-what-you-think-it-means... Excerpt from Karen Hao's Empire of AI: ""The Open in openAI means that everyone should benefit from the fruits of AI after its [sic] built, but it's totally OK to not share the science." "Yup," Musk responded."

"Het is onmogelijk je neutraal te verhouden tot". Wat een opgeblazen vaagtaal.

"Very good that Australia has recognized Palestine, but what next? What are you doing to use your leverage to stop the carnage? Starts with cutting all ties with the state that has been accused of war crimes, crimes against humanity and genocide." - Francesca Albanese www.abc.net.au/listen/progr...

The part of the fascism that scares me the most and terrifies me that we're too late is how successful the technofascists have been in getting people to outsource their critical thinking.

This is some thoughtful analysis.

Hmmm. Universities vs AI. Goodbye universities. Hello alternative facts.

Autistic burnout isn’t a “rough patch”. It’s the bodymind shutting down after years of masking, unmet needs, and constant demand. It can last months or years. It’s not failure – it’s years of survival catching up. Recovery means redesign, not pep talks. It means making changes.

In this paper, @olivia.science Radically Redefines AI as any relationship between humans and artifacts “where it appears as if cognitive labour is offloaded onto such artifacts”. She distinguishes 3 types of relationship: AI that “replaces, enhances, or displaces cognition”. See Table 1 below. 2/n

Labour unions very strong but they are weird pro AI at the moment however @jelmerevers.com is a bit (a lot) higher up and agrees with me and @irisvanrooij.bsky.social and coauthors 🫡

@westbynoreaster.bsky.social long story short, unions negotiated with UNL, which represents universities' top management levels but/and/so is very pro-ai, leading to some dreadful technosolutionist AI talking points in a collective labour agreement, which may or may not be walked back or diluted

Sherry Turkle’s phrasing was “evocative mirror” and I like that one. It always says more about the perceiver than the technology. Folks who think LLMs are smart are indiscriminate, those who think they’re the new Industrial Revolution don’t know anything about the old one, and so on.

Ouija board is the best analogy

Full paper here: zenodo.org/records/1706...

Fitting with the ouija board theme … bsky.app/profile/iris...

Some folks seem to think it is possible to have a pact with the devil and use it to teach about ethical AI.