This is the text of a seminar given at the Goldsmiths Centre for Philosophy and Critical Thought on June 11th 2025
I would like to thank
the Centre for Philosophy and Critical Thought
for inviting me to give this seminar.
This talk is titled
'The role of the University is to resist AI',
and takes as its text
Ivan Illich's 'Tools for Conviviality'.
AI's impact on higher education
come primarily from historical forces,
not from its claim to be
sci-fi tech from the future.
Society can't throw up its hands in shock
as students outsource their thinking
to simulation machines
when fifty years of neoliberalism
has masticated education into something
homogenised, metricised and machinic.
Meanwhile, so-called Ed Tech has claimed for decades
that learning is informational rather than relational
and ripe for technical disruption.
When Illich refers to tools,
he's taking this broader view.
As he writes:
"I use the term 'tool' broadly enough to include
not only simple hardware such as drills, pots, syringes, brooms, building elements, or motors,
and not just large machines like cars or power stations;
I also include among tools productive institutions such as factories
that produce tangible commodities
like corn flakes or electric current,
and productive systems for intangible commodities
such as those which produce
'education,' 'health,' 'knowledge,' or 'decisions'."
I want to ask the question
"What kind of tool is AI?",
to help determine whether Illich's ideas
can assist us in responding to it.
ai
Contemporary AI is a specific mode of
connectionist computation
based on neural networks
and transformer models.
AI is also a tool in Illich's sense;
at the same time, an arrangement of
institutions, investments and claims.
One benefit of listening to industry podcasts,
as I do,
is the openness of the engineers when they admit
that no-one really knows what's going on inside these models.
Let that sink in for a moment:
we're in the midst of a giant social experiment
that pivots around a technology
whose inner workings are unpredictable and opaque.
But there are some things we can be sure of,
which is that the whole show depends on scale.
None of the party tricks
of latent space representations
or next token predictions
will work without tons of data
or wall-to-ceiling computers,
and if you want to beat the other lot
you need more of all of it.
This means that AI is actually a giant material infrastructure
with huge demands for energy, water and concrete,
while the supply chain for specialised computer chips
is entangled with geopolitical conflict.
It also means that the AI industry will beg, borrow and steal,
or basically just steal,
all the text, images and audio
that it can get its spidery hands on.
A marginal point of note for both AI and UK higher education
was a recent outburst
by former politician and Facebook exec Nick Clegg,
who was complaining that copyright is killing the AI industry.
Clegg being, of course, the man who betrayed his promise
to scrap student fees in 2011,
and is now betraying
writers, artists and musicians.
In any case, scale is a core concern for Illich,
and in Tools for Conviviality he writes:
"It is possible to identify a natural scale.
When an enterprise grows beyond a certain point on this scale,
it first frustrates the end for which it was originally designed,
and then rapidly becomes a threat to society itself.
These scales must be identified
and the parameters of human endeavours
within which human life remains viable
must be explored."
higher education
Generative AI's main impact on higher education
has been to cause panic about students cheating,
a panic that diverts attention from
the already immiserated experience
of marketised studenthood.
It's also caused increasing alarm about staff cheating,
via AI marking and feedback,
which again diverts attention from
their experience of relentless and ongoing precaritisation.
The hegemonic narrative calls for universities
to embrace these tools
as a way to revitalise pedagogy,
and because students will need AI skills
in the world of work.
A major flaw with this story
is that the tools don't actually work,
or at least not as claimed.
AI summarisation doesn't summarise;
it simulates a summary based on
the learned parameters of its model.
AI research tools don't research;
they shove a lot of searched-up docs
into the chatbot context
in the hope that will trigger relevancy.
For their part, so-called reasoning models
ramp up inference costs
while confabulating a chain of thought
to cover up their glaring limitations.
The way this technology works means that
generative AI applied to anything
is a form of slopification,
of turning things into slop.
However,
where AI is undoubtedly successful
is as a shock doctrine,
as a way to further precaritise workers
and privatise services.
This casts a different light
on the way OpenAI, Anthropic and Google
are circling higher education,
dangling offers of educational LLM programmes
that have already signed up the LSE,
California State University
and the whole of Estonia's high school system.
It brings to mind Illich's warning about radical monopolies:
"I speak about radical monopoly when one industrial production process
exercises an exclusive control over the satisfaction of a pressing need,
and excludes nonindustrial activities from competition.
The establishment of radical monopoly happens
when people give up their native ability
to do what they can for themselves and for each other,
in exchange for something 'better'
that can be done for them only by a major tool."
critical thought
More specifically, in light of today's seminar
what does this mean for critical thought?
The University of London is already promoting
a tool that provides
"personalised AI generated feedback in under 2 minutes
...including advice on critical thinking".
But thinking for yourself
is a frictional activity
not a statistical correlation.
An AI-mediated essay plan
has already missed the point
by bypassing the student's own capacity
to develop and substantiate propositions
about the world.
When similar AI was adopted by the LA Times
to add journalistic balance to opinion pieces,
it rebalanced an article about the KKK
by clarifying the Klan as
a product of white Protestant culture
that was simply responding to societal changes.
There's already research indicating
students' problem-solving and creativity can decline
when they off-load cognition to chatbots.
Google's recently launched Gemini 2.5 Flash model
even has a "thinking budget"
feature that
allows control over AI's so-called reasoning levels,
and boasts of reducing output costs by up to 600%.
Moreover,
the more these models claim to be safe for education,
the more they become machines for
metapolitical control.
Whatever ketamine-fuelled 3am tweak of the system prompt
made Grok insist on discussing white genocide
will be more powerfully nuanced when done by nice people
who are applying ministry-approved fine tuning.
Critical thought is not
something you can stochastically optimise,
and I agree with Hannah Arendt
that thoughtlessness is a precondition for fascism.
students
But what about the students?
Aren't we doing them a disservice
if we don't prepare them for a world of AI?
As soon as they leave university,
they're going to be faced with AI-powered recruitment apps
that mashup deep learning and psychometrics
to predict their future value to the company.
In their white collar job
they'll use AI to write reports for management
who'll use AI to summarise them,
while every chatbot interaction feeds analytics
that assess their alignment with corporate goals.
They'll constantly be faced by
AI that fails to actually complete the task at hand,
despite the CEO's beliefs to the contrary,
and will have to work overtime
to backfill its failures.
If they're stressed or depressed
they'll be passed to AI-powered therapy bots
optimised for workforce adaptation rather than
for getting to the bottom of their distress.
According to a survey of 16-21 year olds
by the British Standards Institution,
46% said they would rather be young in a world
without the internet altogether.
That's the result of
two decades of algorithmic toxicity;
how long do you think it will take
for them to feel the same about AI?
And yet universities are falling over themselves
to convince faculty and students alike
that AI is the only possible future for higher education,
and research funders only want to fund
things that add AI instead of
researching alternatives.
Any university with a focus on graduate employability
should question the hype about workplace AI
which, in the words of Microsoft's own researchers,
can result in the deterioration of cognitive faculties
and leave workers atrophied and unprepared.
Students already have a sackful of reasons
to be disaffected from the world we're bequeathing them;
do we really want to find out what happens
when we gaslight their doubts about
the value of a synthetic education?
As Illich put it in Tools for Conviviality:
"When ends become subservient to the tools chosen for their sake,
the user first feels frustration
and finally either abstains from their use or goes mad"
Or, as the 17 and 18 year olds
from state schools rated less well
by Ofsted's crappy Covid spreadsheet
put it more succinctly;
"Fuck the algorithm".
labour government
Whatever we or the students might feel
about the role of the university,
our political masters are quite clear
that the only direction of travel is more AI.
This Labour government is possibly
the most AI-pilled in the world,
so at least they got their wish
to be world leading at something.
The AI Action Plan issued in January
is a heady mix of nationalist vibes and startup pitch
that's going to 10x AI,
while handing over land and the electricity grid
to a rash of data centres
in so-called AI Growth Zones.
Labour's single political vision
is growth through AI,
where scaling tech
will somehow stop people voting Reform
or burning down immigrant hostels.
This is a vision articulated by the Tony Blair Institute in reports titled
‘Governing in the Age of AI: A New Model to Transform the State'
and
'The Future of Learning: Delivering Tech-Enabled Quality Education for Britain'.
As an aside,
as we all need a laugh in the face of this nonsense,
their research into how many jobs
would be replaced by AI
included asking ChatGPT.
It does indeed seem that the chef's kiss
in the managerial dismantling of higher education
is going to be from the lips of a chatbot.
What's just as bad
is the way AI is being shoved into other vital services
like there's no tomorrow.
The combination of the Data (Use and Access) Bill
and the Fraud, Error and Recovery Bill
are a literal recipe for repeating Australia's 'robodebt' disaster at scale.
It's like we've learned nothing from the Post Office Horizon IT scandal.
The Department of Work and Pensions is leading the charge
in seeking algorithmic ways to
optimise the disposability of the disabled,
in line with government rhetoric about social burden.
Deep learning has historical and epistemological
connections to eugenics
through its mathematics, its metrics
and through concepts like AGI,
and we shouldn't be surprised if and when
it gets applied in education
to weed out 'useless learners'.
It looks increasingly like the twinning of
the Labour government's fear of Reform UK
and its absolute commitment to AI
are going to bring about the same fusion
of high tech and reactionary politics
as we've seen with MAGA and Silicon Valley.
resistance
I'm proposing that
the role of the university is to resist AI,
that is, to apply rigorous questioning
to the idea that AI is inevitable.
This resistance can be based on environmental sustainability
when looking at AI's carbon-emitting data centres
and their seizure of energy, water and land.
It can be based on the defence of creativity
when looking at the theft of creative work
to train tools that then undermine those professions.
It can be based on decolonial commitments
when looking at AI's outsourcing of exploitative labour
to the global south,
and its dumping of data centres in the midst of deprivation.
Resistance is necessary to preserve
the role of higher education in developing a tolerant society.
For the alternative, we only have to look at the resonances
between right wing narratives
and the ambitions of the tech broligarchy,
resonances that are
antiworker, antidemocratic,
committed to epochal transformation,
resentful and supremacist.
Resonances which, channelled through a UK version of DOGE,
will finish off university autonomy
in the name of national growth and ideological alignment.
DOGE has provided a template
for complete political and cultural rollback,
exploiting AI's brittle affordances
to trash any pretence at social contract.
What the so-called educational offers from AI companies
are actually doing is a form of cyberattack,
building in the pathways for the hacker tactic
of 'privilege escalation'
to be used by future threat actors,
especially those from a hostile administration.
This is why our resistance
needs to be technopolitical.
I'm proposing that higher education
look towards thinkers like Ivan Illich
for an alternative approach
to assessing what kinds of tools
are both pedagogical
and convivial.
illich
Illich proposed what he called counterfoil research
to reverse the kind of obsessive focus on
the refinement of thoughtless mechanism
so visible in the AI industry.
He said that
"Counterfoil research has two major tasks:
to provide guidelines for detecting
the incipient stages of murderous logic in a tool;
and to devise tools and tool-systems
that optimize the balance of life,
thereby maximizing liberty for all."
Illich's purpose in Tools for Conviviality
was "to lay down criteria by which
the manipulation of people for the sake of their tools
can be immediately recognized".
We can take advantage of
subsequent efforts to define
specific starting points,
such as the
Matrix of Convivial Technologies,
which lays out a structured way
for any group developing or adopting a technology
to ask questions about key aspects
such as relatedness (how does it affect relations between people?)
and bio-interaction (how does the tech interact with living organisms and ecologies?).
What we need right now,
instead of more soft soap about responsible AI
or consultancy hype about future jobs,
are institutes that assemble the emerging evidence
of AI's actual consequences
across different material, social & political dimensions.
To stay relevant as spaces for higher education,
universities will need to model
the kind of social determination of technology
which has been buried since the 1970s;
the preemptive examination of tech's value to society.
As Illich says:
"Counterfoil research must clarify and dramatize
the relationship of people to their tools.
It ought to hold constantly before the public
the resources that are available
and the consequences of their use in various ways".
people's councils
While Tools for Conviviality
is a general argument that technology
should be subject to social determination,
and the Matrix of Convivial Technology
gives us a set of specific starting points,
it's pretty clear that the drive to AI
has already advanced from regulatory capture
towards institutional and state capture.
In the UK we already have Palantir
placed at the heart of the NHS,
a military-intelligence company founded by Peter Thiel
that openly espouses cultural supremacy.
The UK exec is the grandson of Oswald Mosley
and, after meeting Keir Starmer,
he said of the Prime Minister
"You could see in his eyes that he gets it".
Instead of waiting for a liberal rules-based order
to magically appear,
we need to find other ways to organise
to put convivial constraints into practice.
I suggest that a workers' or people's council on AI
can be constituted in any context
to carry out the kinds of technosocial inquiry
advocated for by Illich,
that the act of doing so
prefigures the very forms of independent thought
which are undermined by AI's apparatus,
and manifests the kind of
careful, contextual and relational approach
that is erased by AI's normative scaling.
When people's councils on AI are constituted
as staff-student formations
they can mitigate the mutual suspicion engendered by AI.
The councils are means by which to ask rigorous questions
about the conviviality of AI
and, as per Illich's broad definition of tools,
to ask about the conviviality of universities
by applying the same set of criteria
to both infrastructures.
They're also be an opportunity
to form coalitions with allies outside higher education
whose work or lived experience
relates to programmes of study,
and is also being undermined by
degenerative AI,
from the software engineers in DeepMind and Microsoft
concerned about the entanglement of AI with genocide
to the health professionals who see funds diverted
into shiny AI projects instead of fixing the basics.
It's also clear that AI is flooding into
primary and secondary education,
both organically thanks to big tech,
and systemically through government initiatives.
We need practical collaboration
between educators at all levels
to challenge the way AI is flooding the zone,
or the students of the future will be
fully AI-cooked even before they make it to university.
More optimistically, it's not so hard to imagine a near future
where a course or programme that's vocal
about the way it's limited or even eliminated AI
will have additional appeal
as an alternative to the current pathway where
universities conclude that, thanks to AI,
they don't really need most lecturers,
and then students come to the conclusion that,
for similar reasons,
they don't really need the universities.
The function of people's councils on AI
is also to imagine a future for universities
in societies heading for collapse,
where the bridgeheads to a desirable future for all
aren't correlational computations
but campuses and communities.
imagination
I want to conclude by emphasising
that the proposition
that the role of the university is to resist AI
is not simply a defence of pedagogy,
but an affirmation of the social importance
of imagination.
The technopolitical transformation
of which AI is a part
isn't simply a matter of market capture,
but of a wider nihilism
that seizes material and energy resources,
driven by an unrelenting will to power
and the reformulation of racial supremacy
via algorithmically-mediated eugenics.
It's important to talk about resistance
as a way to find resonant struggles
that can amplify each other.
The capacity for resistance
draws on the resources of independent thought
and critical reflection
which are the qualities I've argued
are diluted or dissolved
by a dependence on AI.
These qualities aren't developed solely or mainly
through time at university,
and yet it is also true that students
have often formed a catalytic part
of many social movements.
In some sense, and possibly despite itself,
the university has been a space for developing
collective forms of hope and imagination
which are not only in short supply
but are actively foreclosed
by technogenic patterns of
social and psychic ordering.
The role of the university isn't to
roll over in the face of tall tales about
technological inevitability,
but to model the forms of critical pedagogy
that underpin the social defence
against authoritarianism
and which makes space to reimagine
the other worlds that are still possible.