I was asked to give a workshop on “AI” at a PhD conference (being a professor of “AI, Language Diversity, and Communication Technologies” is asking for trouble, of course). I decided to write back to say the workshop I could offer was one on how not to use it. Abstract:
In a time when it seems impossible to escape talk of the endless “opportunities” and “possibilities” of using “AI”, in this workshop we take a step back and think about how NOT to use it. I will cover a bunch of joyful activities that make the research process meaningful, and I invite participants to think along with me. We will discuss, among other things, the joy of reading deeply and widely; the struggle, and liberation, of finding your own voice in academic writing; and the risks and rewards of slow science (Stengers 2018).
The mirage of “AI” reaches us through a narrative centered on inevitability, productivity, and efficiency. It is therefore at loggerheads with the values of a healthy academia, which questions narratives of inevitability, prioritizes process over product, and realizes the value of friction for creativity. This workshop is for everyone who is bored about the typical AI sales pitch and interested in how to do the best science possible and build a better academia together. The workshop promotes a values-first rather than a tech-first perspective and articulates what you probably already suspected: it is perfectly okay not to use “AI”.
Dingemanse, M., and Gijske de Boo (2023) Human creativity in the age of text generators. The Ideophone doi.org/10.59350/z003g-x2s41
Guest, Olivia, Marcela Suarez, and Iris van Rooij. 2026. ‘Towards Critical Artificial Intelligence Literacies’. In A Research Agenda for Critical AI Studies, edited by Ana Valdivia. Edward Elgar Publishing. https://doi.org/10.5281/zenodo.17786243
Illich, Ivan. 1973. Tools for Conviviality. Open Forum. London: Calder and Boyars.
Stengers, Isabelle. 2018. Another Science Is Possible: A Manifesto for Slow Science. Cambridge, UK: Polity.
Framing the workshop like this was liberating and useful. It was liberating, because it was clear from the start that this was not going to be a middling both sides kind of workshop about the “challenges, but also the opportunities” of AI. (A form of false balance that has corrosive effects on our standards and values.) It was useful, because it allowed us to focus on aspects of the scholarly process that spark joy, which is a key thing that we don’t do often enough.
Too often, workshops are primarily aimed at how to get better at X or more effective at Y. We rarely stop to appreciate the fact that, as PhDs, we have all (i) survived and (ii) escaped an education system built for uniformity and mediocrity, and that it is an incredible privilege to have the freedom, scholarly and otherwise, to work at a public university. The question is what we do with this freedom.
Jumping on the “AI” bandwagon is essentially a capitulation, a careless renouncement of that freedom. It is saying: I am happy to go all in on the productivity and efficiency narrative and to hand in my nascent and hard-won critical thinking skills if that means I can work with machines of mediocrity (Dingemanse 2026) that help me simulate outcomes without understanding. I think considered and principled refusal is a much more interesting, complex, and nuanced option. I say this as someone who studies this technology and its ramifications for human sociality and creativity. But anyway you don’t need to listen to me; you can just educate yourself on the topic of critical AI literacies (Guest et al. 2026) and the impossibility of automating ambiguity (Birhane 2022).
Post-workshop reflections
I devised the workshop around three questions that participants had gotten time to think about beforehand:
- What are you reading?
- What is an example, to you, of excellent writing?
- What aspect of your work would benefit from slowing down?
The reasoning behind the two first question is that for this audience (mostly PhDs and some of their supervisors), a large part of academic work is about writing, and you don’t get to be a good writer without reading a lot. So both activities play a huge role in the process of becoming a skilled scholar. And both activities are of course actively being undermined by the promises of hyped “AI” technologies.
It was interesting and informative to have a conversation about these matters. People differ in their reading habits, in whether they regard reading as ‘for work’ or ‘for pleasure’ or both (some don’t really make that distinction), in what, where and when they read, and there’s something very useful in bringing out the variety in people’s experiences and preferences. The quest for excellent writing also brought out some wonderful tips —Ted Chiang, Daphne du Maurier and Ursula Le Guin among them— that helped us see, collectively, how reading anything, fiction included, can help broaden and strengthen our own (non-fiction) academic writing.
I introduced the question about slowing down by reference to Isabelle Stengers’ manifesto for slow science. Collectively, the audience members listed a great many things that would benefit from slowing down. They said they would love to find time to slow down
- for building social relations
- for diving into empirical data together
- to talk to each other and exchange perspectives
- to revisit and revise our writing
- to learn from each other by reading and commenting on each other’s work
- to avoid comparing oneself to others
I can’t speak for the participants, but at least I myself found it immensely valuable to sit down, think about what makes our work worth it, and articulate some of these things together. These are activities that scholars, and especially junior scholars, realise they need time for in order to do good work. They are, of course, also exactly the activities that the “AI” hype takes away from us with its promises of relentless speed, productivity and efficiency. I hope that for at least some of the participants, it’s been useful to take a step back and realise that another science is possible.
I introduced participants to the pentagram of critical AI literacies from Guest et al. 2026, and we saw that many of the shared values and worthwhile activities were precisely those that foster conceptual clarity, critical thinking, decoloniality, respect for expertise, and slow science.
A final and very justified question from a PhD participant was who is to be responsible for slowing down. We all know people, this person said, who go wild on “AI” and who seem to be cranking out study after study. To this I said two things. First: That it is the responsibility of seniors in the field to speak out against harmful productivity myths and create room for truly good research. If we don’t, there won’t be a field left to speak for: we’ll be reduced to “AI agents” reading slop and writing slop and reviewing slop. (There is an interesting Ted Chiang story about this.) Second: That what goes up must come down. Amidst a rising tide of slop science, one thing we know is that true artisanal high quality scholarly work is only becoming more scarce, and therefore more valuable. As I wrote in 2020 (with the first GPT around the corner): Counterintuitively, there was never a better time to be a scholar.