Whom Is the University For? Artificial Intelligence, Cognitive Surrender and the Class of Thinking

Whom Is the University For? Artificial Intelligence, Cognitive Surrender and the Class of Thinking

A comradely reply to Can Ertuna's piece in BirGün, "Artificial intelligence, cognitive surrender and the university" (in Turkish), and a draft on the ideal university

First, a Short Summary

• What does the piece say? Can Ertuna argues that the "don't go to university, four years is too much" discourse, running from Sam Altman to Alphan Manas, is a demand of capital, and that the university is not a "job factory" but an institution that produces knowledge and teaches critical thinking. According to a study based on the data of more than four million students, students are handing their mental effort over to AI ("cognitive surrender"): while success in unsupervised assignments rises, the rate of correct answers in supervised exams has fallen from 80% to 60%.
• Where it is right: The "there is no need for university" discourse is not a neutral forecast but a class programme. Its defence of the basic sciences and the social sciences is also well placed.
• What could be developed: Cognitive surrender is not individual laziness or "the bad effect of technology". In an order where the diploma has become a commodity and the grade an exchange value, and the student has to both work and study, the shortcut is rational. The root of the problem lies not in AI but in the commodification of education.
• A second point that could be developed: The piece somewhat idealises the university as an institution that "preserves and transmits accumulated knowledge". The bourgeois university was never an island of enlightenment above classes. And in Turkey, what made the university what it is today is not only the market but also the YÖK of 12 September, the dismissals by emergency decree (KHK) and appointed rectors.
• The real danger: AI companies are pulling the knowledge the university produces (articles, theses, lecture notes) as training data, then renting it back to the university as an institutional licence. The university becomes both supplier of raw material and customer. This is the expropriation of the general intellect.
• The solution is not bans and surveillance: Returning to in-class exams, oral exams and proctoring software suppresses the symptom and enlarges the digital panopticon. The solution is to break the exchange value of the grade and the diploma, and to tie learning back to use value.
• The ideal university: Free of charge, open to everyone and for life; democratically governed; producing and sharing knowledge as a commons; with public computing and open model infrastructure; uniting mental and manual labour (polytechnic); connected to society's problems and to working people's organisations.
• The socialist university: If the ideal university is a goal to be won through struggle within capitalism, the socialist university is a step beyond it: an institution where the private ownership of knowledge is abolished, the division between mental and manual labour is overcome, and, as working time shrinks, society as a whole becomes a student. Learning from both the achievements and the mistakes (dogmatism, party-controlled science, hostility towards intellectuals) of the Soviet workers' faculties and the Cuban and GDR experiences.
• The answer: In the age of AI we need the university more. But not today's university. A university that does not hand thinking over to the monopolies, and that protects society's common intelligence.

Now let us go into the details.

---

Dear Young Comrades,

Some of you reading this piece are sitting in a lecture hall right now; some are trying to finish a shift in a café, a courier depot or a call centre to scrape together the registration fee; and some are asking themselves, "what's the point of studying?" On your screen, on one side, are billionaires saying "AI will do everything, don't waste four years", and on the other a chatbot that writes your assignment. Your lecturer is trying to work out whether the bot wrote the assignment, and you are trying to make sure no one can tell that it did.

In this scene everyone seems to be fooling everyone else. Yet the real question is different: Who set up this game, and whom does it serve?

Can Ertuna's piece, published in BirGün on 21 October, opens an important door onto this question. Let us go through that door and walk a little further.

---

What Does the Piece Say?

The skeleton of Ertuna's piece is as follows:

The debate on the purpose of the university. Those who see the university as an "education factory" that trains a workforce propose reshaping education around technology in the face of AI. For the author, by contrast, the university is an institution that preserves and transmits accumulated knowledge and produces new knowledge.

The "don't go to university" discourse. It recounts Sam Altman dropping out of Stanford and saying that a four-year university is too much, and Alphan Manas advising his children not to go to university or to go only briefly. The author says this discourse is connected to the demands of capital and the market.

Dark Academia. Citing Peter Fleming's book Dark Academia, it criticises the transformation of universities into "job factories" serving the demands of industry.

The value of the basic sciences. It argues that the practical use of mathematics is not immediately visible but forms the basis of new technologies in the long run, giving the spin-offs of the Apollo programme as an example. It relays the warning of Giordano Scarciotti of Imperial College: if young people turn away from the field because "AI will put mathematicians out of work", the number of people able to understand advanced mathematics could fall sharply within twenty years.

Cognitive surrender. A study (June 2026) by researchers from UC Irvine and McGraw Hill examined data from late 2022 to 2026 for more than four million students a year on the ALEKS mathematics platform. What the researchers call "cognitive surrender": instead of solving the problem themselves, students hand the mental effort over to AI. The result is striking: success rises in unsupervised exercises, while the rate of correct answers in supervised exams falls from 80% to 60%.

The institutions' response. It notes that universities such as Berkeley, Stanford and Chicago have banned or restricted the use of AI in some courses (especially in the social sciences and law), and that there has been a return to in-class written exams and oral presentations.

Conclusion. AI will become part of life, but thinking must not be handed over to it. To be able to see errors and nuances, both technical knowledge and the social sciences are needed. The future must not be left to a handful of tech oligarchs, business leaders and political elites.

---

Where the Piece Is Right

Let us give it its due first, because in its main orientation the piece stands on the right side.

First: It diagnoses the "there is no need for university" discourse as an ideology. This matters. The forecasts that come out of the mouths of tech bosses are presented to us like a weather report: "It's going to rain, take your umbrella." Yet these are not forecasts but wishes. And they are the ones who will make it rain.

Second: It defends the basic sciences. The short-term horizon of the profit calculation counts mathematics, philosophy and history as "useless". But what we call AI today itself rests on a hundred years of "useless" mathematics, probability theory, linear algebra and statistics. It would have been good if no one had asked those working on the mathematics of neural networks in the 1940s, "what is the market value of this?" They did ask; that is why the field went through decades of "winter".

Third: It stresses the necessity of the social sciences. To see what a text produced by a model conceals, whose point of view it takes as natural and which silences it reproduces, we need history, sociology and philosophy. Even calling this "nuance" is a little modest; the issue is ideology.

Fourth: It says the future must not be left to a handful of oligarchs. True. But we also need to name where these oligarchs come from, and which order they are the product of. We will get to that.

---

A Class Reading: What the Piece Could Develop

"Don't go to university": whose advice, and to whom?

Sam Altman dropped out of Stanford. But first he got into Stanford. This detail is not a small one.

The stories of billionaires who dropped out of university (Gates, Zuckerberg, Altman) always have the same structure: a young man from a privileged family, who got into an elite school and pocketed the network, the access to capital and the safety net it offered, saying "I dropped out". Dropping out is the luxury of those who can come back.

So where do these billionaires' own children go? Silicon Valley's elite families send their children to screen-free private schools with human teachers and small classes. Places at elite universities are still fought over, and their fees are still rising. In other words, the university remains for the elite; for the poor, "a short course, a certificate and a chatbot" are recommended.

Here is the class picture: in the age of AI, education is splitting in two.

• For the upper classes: human teachers, small classes, discussion, face-to-face relationships, broad culture, learning to think.
• For the lower classes: automated platforms, screens, isolated drills under the name of "personalised learning", short on-the-job skills training, learning to give prompts.

Alphan Manas's advice not to go to university may be an option for an upper-middle-class family in Turkey. But for the child of a working-class family in Diyarbakır, Esenyurt or Ereğli, the university, despite all its decay, is still the last door to climbing out of their class, or at least to not being condemned to the factory or to courier work. Closing that door because "there is AI anyway" means telling these young people, "you will be the ones who receive commands, not the ones who give them."

Cognitive surrender: laziness, or a rational response?

The study's finding is important: success rises in unsupervised practice and falls in supervised exams. In other words, students look more "successful" but learn less.

It is easy to read this as "young people have grown lazy; technology is rotting their brains". But that is a reading that puts the problem on the individual and does not see class. Let us ask: Why does the student choose the shortcut?

Because education has become a commodity. In Marx's language, a commodity has two faces: use value and exchange value. The use value of knowledge is to understand it, to grasp the world with it and to change the world. The exchange value of the diploma and the grade is that they are an entry ticket to the labour market.

The student learns very early that what is demanded of them is not knowledge but a grade, not understanding but a certificate. Exam anxiety, loan debt, KYK (state student loan) repayments, part-time work, long commutes, expensive dormitories... In these conditions, a young person for whom time is the scarcest resource turning to a tool that produces the grade in the cheapest way is not laziness but economic rationality. The system asks them to produce a commodity; they use the most efficient means of production.

Here is a textbook example of alienation. Like the worker Marx describes in the 1844 Manuscripts, the student:

• is alienated from their product: the assignment they hand in is not theirs.
• is alienated from the process: they do not expend the labour of thinking; they hand it over.
• is alienated from their species-being: they move away from the conscious, creative activity that makes us human, that is, from thinking.
• is alienated from other people: instead of working together with classmates, everyone is alone on their own screen, with their own bot.

So "cognitive surrender" is not a disease but a symptom. The disease is the commodification of education. AI did not create this disease; it made it visible and fast.

Mental Taylorism: A new division within mental labour

In Labor and Monopoly Capital (1974), Harry Braverman describes the essence of Taylorism like this: knowledge of the production process is taken from the worker and concentrated in management; the worker is left only with fragmented execution, without understanding. Head and hand are separated, and the head passes over to the side of capital.

In the age of AI this separation goes a step further: a division within mental labour itself. Functions such as understanding, reasoning and synthesis are being moved into models owned by capital. The human is left with fragmented tasks such as "writing prompts", "checking the output" and "correcting and submitting".

The university's role in this process is critical. If the university teaches its students only "to use tools efficiently", it is in fact training the workers of mental Taylorism: knowledge workers without understanding, who only execute, who do not know the logic of their own work, and who can therefore easily be replaced.

The fall from 80% to 60% in the study is exactly this: deskilling. And deskilling is always a loss of bargaining power. A worker who does not understand the work they do cannot defend the value of that work either.

Let us not forget that the same process is happening to academics. Course content uploaded to platforms, automated assessment systems, "efficiency" metrics, an army of hourly-paid and contract academics... Academic labour, too, is being fragmented, made precarious and surveilled.

The expropriation of the general intellect: The university as both mine and customer

There is a point that the piece does not touch on at all, but which is central for us.

In the "Fragment on Machines" in the Grundrisse, Marx speaks of the concept of the general intellect: society's accumulated knowledge, science, common intelligence. In his view, capital objectifies this social knowledge in machines and appropriates it.

This is exactly what is happening today. What were large language models trained on?

• Articles, theses and textbooks written in universities,
• Research shared through open access,
• Knowledge that academics and students shared for free on forums, in encyclopaedias and in code repositories.

In other words, the models were trained largely on the public labour of the university and of society. Then what happened? The same companies began selling these models back to universities under names such as "education edition" and "institutional licence". The university pays rent to use a tool distilled from the knowledge it produced itself.

This is digital enclosure. Just as the common pastures of peasants in England were fenced in and turned into private property, today humanity's common store of knowledge is being shut inside model weights and turned into private property.

The university loses twice in this relationship: once as a supplier of raw material (its knowledge taken without payment), and once as a customer (buying the processed form of that knowledge). A call to "defend the university" becomes much stronger when it includes this dimension too.

The nostalgia trap: Which university are we defending?

Ertuna describes the university as an institution that "preserves and transmits accumulated knowledge". This description is correct; it becomes even stronger when taken together with its historical dimension.

The Humboldtian research university, the Oxford-Cambridge model and the American state university alike were founded according to the needs of a particular class and a particular state: to train bureaucrats, officers, engineers, lawyers and clergy; to supply science to national industry and the army. Throughout history, a significant part of the knowledge produced by the university has been intertwined with the war industry. The Apollo programme that Ertuna gives as an example was itself part of the Cold War and the missile race.

The periods when the university was an "autonomous island of enlightenment" are in fact the periods when social struggles forced it to be: the student movements of 1968, the public expansion that allowed working people's children to enter the university, the fights waged for academic freedom. So what we defend should be not "the good old university" but the public and critical university won through struggle.

For Turkey this is even clearer. It was not only the market that shaped today's university:

• 12 September and YÖK (1981): A coup institution that put the university under central control and abolished its autonomy. Hundreds of academics were expelled from the university under martial law decree No. 1402.
• The KHK dismissals after 2016: Thousands of academics were dismissed, above all the academics who signed for peace; we have now left a decade of this behind us.
• Appointed rectors: Rectors coming by appointment rather than by election; the Boğaziçi resistance was one of the longest-running struggles against this imposition.
• Fee-paying education and foundation universities: Fees rise every year; this year again YÖK had to send foundation universities a warning setting a ceiling on increases. On one side, "a university in every province" with no research infrastructure at all; on the other, fee-paying schools that only those with money can attend.

So in Turkey, what threatens the university is not only tech billionaires; it is the alliance of the authoritarian state and the market. AI is giving this alliance a new tool: both cutting costs (fewer lecturers, more platforms) and tightening control (surveillance software, automated detection of "inappropriate content").

Not "techno-oligarchy" but monopoly capitalism

Ertuna wants the future not to be left to "a handful of tech oligarchs". We agree, but let us sharpen the naming. These oligarchs are not a new feudal nobility that has come from outside capitalism. They are today's most concentrated form of monopoly capitalism: a few companies that own computing power, data, chips and models.

Naming this correctly matters, because it determines the solution. If the problem were "a few ill-intentioned rich people", it would be enough to regulate them, tax them and have them write "ethical principles". But if the problem is a property relation, then the solution is to change the property relation: public and common ownership of knowledge, data and models.

Bans and surveillance: Suppressing the symptom

The institutional responses reported in the piece (bans, in-class exams, oral presentations) are understandable but limited. Some of them are pedagogically valuable too: face-to-face discussion, oral defence and thinking in handwriting are good things.

But note: these responses have another face. The digital panopticon. Many universities have turned to proctoring software that monitors the student's camera, microphone, screen and eye movements during online exams. "AI detection" tools, meanwhile, often give false alarms, and those most harmed by these false alarms are students whose first language is not the language of the exam, and poor and disadvantaged students.

In other words, responding to the problem of "cheating with AI" with surveillance turns the student into a suspect and education into a security operation. Like Bentham's prison design: a person who knows they are being watched all the time disciplines themselves. They do not learn; they conform.

If the root of the problem is the exchange value of the grade, the solution must be sought there too: tying learning back to use value. We will come below to how this could be done.

---

Two Universities at a Glance

The University of the Market and of Authority	The Common and Public University
Education is a commodity; the student is a customer	Education is a right; the student is a producer of knowledge
Aim: people "fit" for the labour market	Aim: people who understand the world and can change it
Measure: grades, rankings, graduate salaries	Measure: understanding, common production, social contribution
Governance: appointed rector, YÖK, board of trustees	Governance: elected bodies, participation of students, academics and staff
Academic labour: hourly-paid, on contract, precarious	Academic labour: permanent, secure, unionised
AI: a closed model rented from a company	AI: an open, auditable model on public infrastructure
Knowledge: behind the walls of publishers and platforms	Knowledge: open access, open data, open source
Response to cheating: bans, surveillance, detection software	Response to cheating: removing the pressure of grades, process-focused learning
Basic sciences: departments closed as "unprofitable"	Basic sciences: a long-term social investment
A campus closed to society	A campus open to the neighbourhood, the factory, the union, the city
Fee-paying, four years, once	Free of charge, open to all, for life

---

The Ideal University: What Should It Be?

We are not drawing a utopia here. Each of these either exists in part somewhere in the world today, or is something won through struggle in the past and then taken back.

Free of charge, open to everyone and for life

In the age of AI, the "study once and be done" model is already collapsing. Capital loads this onto the worker under the name of "lifelong learning": update your skills with your own money, in your own time, at your own risk; if you cannot, it is your fault.

We must turn this around: lifelong learning is not a burden but a right. A 45-year-old accountant, a 30-year-old courier, a 60-year-old retired worker should be able to enter and leave the university. Free associate degree and certificate programmes without entrance exams, evening and weekend classes, educational leave at work. The productivity gains created by technology should go not to unemployment but to shorter working hours, and to devoting the time this frees up to learning.

Democratic governance

Rectors, deans and heads of department should be elected by academics, students and university staff. Central control bodies such as YÖK should be abolished and replaced by democratic coordination bodies formed by the universities among themselves. The budget should be transparent. Students' say should be not decoration but decision-making power.

This is not only a matter of "democracy" but also a matter of knowledge: an institution governed from above cannot produce critical thought. An academic who cannot elect their rector thinks twice before writing an article that criticises those in power.

Producing and sharing knowledge as a commons

All knowledge produced with public money should be open to the public:

• All articles, theses and lecture notes should be open access.
• Software written at the university should be open source.
• Research data should be open data, provided personal data is protected.
• And most importantly: this public knowledge must not be used without payment to train the closed models of private companies. A model trained on public knowledge should be public.

This last point is the counterpart of the "knowledge commons" idea in the age of AI. Openness cannot be one-way. If we leave things open and they enclose them, that is not openness but plunder.

Public computing and open model infrastructure

Today the greatest weakness of academia is this: the computing power to train frontier models exists only in a few companies. This means that the research agenda is set by those companies too. The university researcher has to make do with "testing" the closed model offered by the company from the outside; they cannot look inside it, cannot see its data, cannot replicate its training. This is not science; it is a customer satisfaction survey.

The ideal university should have public computing infrastructure through an inter-university consortium. It should develop open-weight, open-data, auditable models and be able to run them locally. Students' data should not go to a foreign company's servers. This is also a matter of data sovereignty.

Polytechnic: Reuniting head and hand

In Capital and in the Critique of the Gotha Programme, Marx speaks of an education for working people's children that combines mental education, physical education and "polytechnic education", that is, grasping the scientific foundations and the tools of production. The aim is to close the gulf between mental and manual labour.

In the age of AI this idea is topical once again. Because the antidote to mental Taylorism is a person who grasps the whole of their work. A software developer should know not only how to write code but within which social relation that code runs; a doctor not only how to use a diagnostic tool but with what data, whose labour and which biases that tool was trained. An engineer should work in the workshop, a sociologist in the field, a mathematician inside a real problem.

The polytechnic university teaches not a profession but production itself. That is why it is neither a narrow vocational school nor an ivory tower that has turned its back on society.

An institution connected to society

Gramsci says that "all men are intellectuals, but not all men have in society the function of intellectuals." The university's task is to open up space for "organic intellectuals" to emerge from every section of society: people who grasp the problems of their own class scientifically and give voice to them.

For this, the university must go beyond the walls of the campus. It should carry out joint research with unions, professional chambers, cooperatives, municipalities and neighbourhood organisations. Workers' health, earthquake safety, urban transformation, climate, food, energy, algorithmic management... The research agenda should be set not by corporate sponsorship but by the needs of society.

Secure academic labour

The ideal university does not turn its academics into an "academic precariat" of hourly-paid and contract staff living every semester in fear of losing their jobs. Permanent posts, security, the right to organise in a union, time set aside for research. A precarious academic cannot think critically, because criticism means risk.

---

How Should We Teach? Pedagogy in the Age of AI

Effort is a commons: Preserving difficulty

Learning is laborious by nature. Educational scientists call this "desirable difficulties": trying to remember, making mistakes, going down the wrong path and coming back are learning itself. The fifteen minutes of effort you spend trying to solve a problem yourself is your brain making that knowledge your own.

AI's "frictionless" promise is precisely to remove this labour. But when you remove the labour, you remove the learning as well. The finding of the study shows this.

That is why the ideal pedagogy preserves difficulty not with bans but with meaning. Students want to solve a problem themselves because that problem is meaningful to them, because it gives them something other than a grade. This brings us to the next point.

Removing the pressure of grades

If what feeds the shortcut is the pressure of grades and rankings, that pressure needs to be reduced:

• Process-focused assessment: Not only the result but the thinking process itself (drafts, notes, mistakes, corrections) should be assessed.
• Portfolio: The whole body of work a student produces over the semester.
• Oral defence and discussion: Not as a surveillance tool, but as a practice of putting thinking into words.
• Mainly pass/fail systems: Especially in the first years, assessment that reduces competition over rankings and makes room for learning.
• And of course: Social arrangements that reduce the weight of the diploma in the labour market. This is not something the university can solve on its own; it has to do with the labour market and class relations. But it should be voiced as a demand.

Teaching how to open AI up, not just how to use it

"AI literacy" today mostly means "using tools efficiently". That is companies' user training, not the task of the university.

The university should open AI up: it should teach how a model works, what data it was trained on, whose labour that data comes from, the low-paid workers labelling data in Kenya, the Philippines and India, which biases the model reproduces, what energy and water it consumes, and whose property it is. This is critical AI literacy. This is where technical knowledge and social science meet.

To use a tool without understanding it is to be used by it.

Collective learning

Learning with AI is mostly isolating: one student, one screen, one bot. Yet human beings, as social beings, learn best together with others.

The ideal pedagogy puts joint projects, discussion groups, peer teaching and joint research at the centre. A group of students jointly researching the earthquake risk of a neighbourhood, the accidents at a workplace or a platform's algorithmic pay system produces knowledge, solidarity and the experience of organising all at once.

Basic sciences, social sciences, craft

Scarciotti's warning is well placed: if young people move away from mathematics because "the machine does it anyway", in twenty years there will be no one left to understand, oversee and correct the machine. Then whatever the machine's owner says will be true.

That is why mathematics, physics, philosophy, history, languages and the arts are not a "luxury" but a precondition of democratic oversight. If a society cannot raise citizens who understand the technology they use, it surrenders to the owners of that technology. If the individual form of cognitive surrender is a student having a bot write their assignment, its social form is a country handing its capacity to think over to a few companies.

Small classes, many teachers

Where should the productivity gains AI provides go? Capital's answer is clear: fewer lecturers, larger classes, more platforms. Our answer should be the opposite: let the time technology saves go to reducing the number of students per teacher. Classes of twenty, one-to-one tutoring, face-to-face discussion. Let the education offered to the children of the elite be offered to everyone.

---

The Place of the University in the Modern World

So how much do we need the university in the age of AI?

Our answer: more than ever before. But not in its present form.

Let us think about this: the production of knowledge is shifting more and more to private laboratories. The biggest AI research is no longer done in universities but behind the closed doors of a few companies. Findings are not published, or are published as marketing copy. Independent verification is impossible. Openness and reproducibility, the basic principles of science, are suspended on the grounds of trade secrecy.

In these conditions the university is the last stronghold of three things:

• Society's memory: the place where accumulated knowledge is preserved independently of companies' balance-sheet cycles.
• Society's critique: the only institutional space able to examine technology, power and capital independently (provided we fight to protect this independence).
• The reproduction of society's common intelligence: the place where new generations learn to think, to understand and to question.

If these three functions disappear, what remains is this: a few companies that think, and millions who give commands. Or more accurately, millions who think they are giving commands but are in fact receiving them.

That is why "is there a need for university?" is the wrong question. The right question is: Whose university? Whose knowledge? Whose intelligence?

---

What Should a Socialist University Be Like?

The "ideal university" we have described so far is in fact a set of goals that can be won through struggle within capitalism. Free education, elected rectors, open access, secure academics... Each of these is a demand that can be defended today in the street, in the lecture hall and in the union. But let us be honest: in a society ruled by capital, these gains are always open to being taken back. We saw this at Boğaziçi, with the 1402 dismissals, with the KHKs. As long as the private ownership of knowledge, the logic of profit and the class character of the state persist, the public university remains an island, and the waves constantly erode that island.

So we must also ask this question: What would the university be like in a society where the rule of capital has been abolished?

First, let us learn from history: The achievements

The socialist university is not a blank page. The experiences of the twentieth century leave us both inspiration and warnings.

The Soviet workers' faculties (Rabfak). In 1919, in the middle of the civil war, Soviet power set up "workers' faculties" attached to the universities. The aim was for workers and peasants who had not received secondary education to go through preparatory education and enter higher education. Within a few years, hundreds of thousands of working people's children entered institutions that under tsarism only nobles and the bourgeoisie could enter. They became engineers, doctors, teachers and scientists. This was the first time in history that the class door of the university was broken open on such a scale.

Cuba: "Let the university be painted the colour of the people." At the end of 1959, receiving an honorary doctorate at the Central University of Las Villas, Che Guevara told the university that it must paint itself black, mulatto, worker and peasant; not only its students but its teachers too. With the university reform of 1962, education became free of charge, and together with the literacy campaign, higher education was opened to the people. In later years, university education was spread out to the municipalities, so that classes were given in every town. A small island country under embargo built a medical school that trains doctors for the world and a biotechnology sector that develops its own vaccines.

The GDR: Workers' and Peasants' Faculties. East Germany, too, set up similar faculties from 1949 onwards; the share of workers' and peasants' children in higher education multiplied within a few years. Combining production and education, students working in enterprises periodically, and the polytechnic school system were parts of this experience.

The common lesson is this: socialism is socialist to the extent that it can open the door of the university to working people.

Let us learn from the mistakes too

The same history also leaves us grave mistakes. A design for a "socialist university" that ignores them is doomed to repeat them.

Party-controlled science: the Lysenko affair. In the Soviet Union from the 1930s to the 1960s, Trofim Lysenko's views on agricultural biology, which had no scientific basis, were declared the "official truth" with political backing; genetics was branded "bourgeois science", geneticists were dismissed, some were arrested, and Nikolai Vavilov died in prison. Soviet biology fell behind for decades as a result. The lesson is clear: scientific truth cannot be determined by a committee decision. Society can democratically determine the priorities of research, but it cannot determine the results of research.

Dogmatism and official ideology. When Marxism itself ceased to be a method of critique and turned into a catechism to be memorised, the social sciences became sterile. An unquestionable course in "scientific socialism" is no less dogmatic than an unquestionable course in market economics. Marx's favourite maxim was "doubt everything".

Hostility towards intellectuals. In China, during the years of the Cultural Revolution, universities were closed for years, academics were humiliated and mental labour was regarded with suspicion. The claim to close the gulf between mental and manual labour turned into the practice of denying mental labour altogether. The lesson: overcoming the division means not abolishing the intellectual, but making everyone an intellectual.

Loss of bureaucratic autonomy. In many experiences, the university turned into a sub-unit of the central planning bureaucracy; student and academic self-management remained on paper. Replacing capital's YÖK with the party's YÖK does not solve the problem.

The principles of the socialist university

In the light of these achievements and mistakes, we can set out the basic principles of a socialist university as follows:

There is no private ownership of knowledge. Patents, copyright and trade secrets are the legal forms of the enclosure of knowledge. In socialist society, scientific knowledge, software, model weights and data are the common property of society. An AI model belongs to humanity because it was produced from humanity's common store. The university is both the producer and the guardian of these commons.

Society becoming a university. Marx says that under socialism the measure of wealth will be not working time but "disposable time", that is, the time people can devote to developing themselves. In the hands of capital, AI and automation mean unemployment and intensification; under social ownership they mean a radical shortening of the working day. The time freed by a shorter working day is devoted to learning, research and art. Then the university ceases to be an institution where young people spend four years; it becomes a space that everyone enters and leaves throughout their lives, where society itself turns into a learning community. Being in the factory in the morning, in the laboratory in the afternoon and at a philosophy seminar in the evening becomes not a privilege but ordinary life.

Overcoming the division between mental and manual labour. The engineer knows the production line; the worker knows the design table. The doctor spends time at the community health centre, the sociologist in the neighbourhood, the software developer in the workplace where their code is used. Knowledge of production is not the monopoly of management but the common knowledge of the producers. The exact opposite of mental Taylorism: every worker grasps the whole of their labour.

Democratic planning, autonomous science. Through workers' councils and democratic planning bodies, society debates and decides what to prioritise: cancer research, earthquake safety, or climate adaptation? But the method and results of research are left to the open, critical, peer-reviewed debate of the scientific community. This is where the lesson of Lysenko lies. Planning sets the priorities; evidence sets the truth.

Self-management. The university is governed by recallable representatives elected by academics, students and staff. These representatives coordinate with the workplace and neighbourhood councils around them. Neither capital's board of trustees nor the party's appointed rector.

The guarantee of criticism. The socialist university must be able to criticise socialist power too. This is not a favour but a condition of the system's health. A society that cannot see its mistakes cannot correct them. Different schools, rival theories and fierce debates are the natural state of the university.

Public and open technology infrastructure. Computing power, data centres, AI models and the infrastructure of scientific publishing are under social ownership and open to all. AI is not a black box that replaces thinking, but a tool with open code, known data and under social oversight. Students can look inside the model, change it and retrain it according to the needs of their own community.

The withering away of the exchange value of the grade and the diploma. As the labour market and competition disappear, the diploma ceases to be an "entry ticket". No one cheats to get a grade, because there is nothing to buy with a grade. Learning returns to use value, that is, to understanding, producing and creating. The material basis of cognitive surrender thus disappears.

Internationalism. Knowledge knows no borders. The socialist university opens its doors to students and scientists from poor countries; like Cuba's Latin American School of Medicine, it shares knowledge not as an export product but as a tool of solidarity.

From today to tomorrow: The bridge between the ideal and the socialist

A Demand of Struggle Within Capitalism	Its Counterpart in the Socialist University
Free education	Society as a whole becoming a student for life
Elected rectors, abolition of YÖK	Self-management through recallable representatives
Open access, open source	Abolition of the private ownership of knowledge
Public computing infrastructure	Social ownership of the whole technology infrastructure
Secure academic labour	Overcoming the division between mental and manual labour
A ban on proctoring software	The withering away of the exchange value of the grade
Joint research with unions and communities	Science tied to democratic planning, autonomous in method
The demand for shorter working hours	Disposable time as the measure of wealth

This table tells us that every struggle we wage today is a brick of tomorrow's university. The demand for free education is the first step of society becoming a university. The struggle for open access is the first step of the socialisation of knowledge. The demand to elect the rector is the first step of self-management.

There is no Great Wall of China between reform and revolution; but to settle for reform is to hand our gains over to capital's next attack.

---

Concrete Tasks

For students

• Not rejecting AI, but not surrendering to it either: try the assignment yourself first, use the tool like a conversation partner where you get stuck, and in the end make sure that what is written is yours.
• Set up reading and discussion groups. Learning together is both more effective and more liberating than learning alone with a bot.
• Object, in an organised way, to proctoring software, to the arbitrary use of "AI detection" tools and to your data being passed to companies. Ask which of your data goes where.
• Do not separate the struggle against fee-paying education, KYK debt and the dormitory problem from this debate. One cause of cognitive surrender is time poverty.

For academics

• Open your course design to process-focused assessment, collective projects and field-based research.
• Question your institution's licence agreements with AI companies: where do student and academic data go? Are the contracts transparent?
• Work to make your publications and teaching materials open access; but demand licences and institutional policies against their unauthorised use in training closed models.
• Strengthen union organising, above all in Eğitim Sen (the Education and Science Workers' Union). The precarity of hourly-paid and contract academics is one of the most concrete obstacles to a critical university.

For IT workers

• Develop open-source education tools that can be run locally, together with universities.
• Voice the demand for public computing infrastructure and open models through your professional organisations and unions such as BilişimSen.
• If you work at an education technology company, do not hesitate to share with the public the surveillance and data collection dimensions of the tools you build.

Public demands

• Free higher education, open to all and for life.
• The abolition of YÖK, the election of rectors, the reinstatement of dismissed academics.
• Open access to all knowledge produced with public money; models trained on public knowledge to be public.
• Inter-university public computing infrastructure and a programme for developing open models.
• A ban on surveillance software in education and on the commercial use of student data.
• Preventing the closure of basic science and social science departments on the grounds of "no demand"; scholarships and posts to support these fields.

---

We Will Not Hand Over Thinking

Dear Young Comrades,

The story you are being told is this: machines will think, and you will tell them what to think. You will be comfortable. But history teaches us something else: whoever holds knowledge holds power. A worker who understands the workbench is always stronger than a worker whose craft has been taken away. A generation whose thinking is taken away cannot debate its own future either.

Can Ertuna is right: thinking must not be handed over to AI. We add one more step: thinking must never be handed over to the owners of that AI. And we can ensure this not through individual discipline, through the will to "restrain myself and not use the bot", but through organised struggle: for free education, for a democratic university, for the common ownership of knowledge.

The university is not a building but a relationship. A relationship in which a teacher passes knowledge to a student, a student to a student, a generation to a generation, and in which knowledge is produced together. No one can rent this relationship to us.

Let us meet in lecture halls, libraries, reading groups and unions. Let us keep thinking together, out loud, stubbornly.

Knowledge is a commons. So is the mind.

---

Sources

• Can Ertuna, "Yapay zekâ, bilişsel teslimiyet ve üniversite" (Artificial intelligence, cognitive surrender and the university), BirGün, 21 October 2026 (in Turkish).
• The study by researchers from UC Irvine and McGraw Hill on ALEKS platform data (June 2026); the findings are based on the account given in Ertuna's piece.
• Peter Fleming, Dark Academia: How Universities Die, Pluto Press, 2021.
• Karl Marx, Economic and Philosophic Manuscripts of 1844.
• Karl Marx, Grundrisse, "Fragment on Machines".
• Karl Marx, Capital, Volume I (the chapters on large-scale industry and polytechnic education); Critique of the Gotha Programme.
• Harry Braverman, Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century, 1974.
• Antonio Gramsci, Prison Notebooks (the notes on intellectuals and education).
• Paulo Freire, Pedagogy of the Oppressed.
• Michel Foucault, Discipline and Punish (on the panopticon).
• Ernesto Che Guevara, speech on receiving an honorary doctorate at the Central University of Las Villas, 28 December 1959.
• Studies in the history of education on the Soviet workers' faculties (Rabfak, 1919) and the GDR Workers' and Peasants' Faculties (Arbeiter-und-Bauern-Fakultäten, 1949).
• On the Lysenko affair and Nikolai Vavilov: Zhores Medvedev, The Rise and Fall of T. D. Lysenko.
• On YÖK's warning to foundation universities about tuition fee increases: BirGün (in Turkish), https://www.birgun.net/haber/yok-ten-vakif-universiteleri-zam-uyarisi-artis-siniri-belirtildi-726756
• Knowledge Commons, "Our Education Dossier Is Out: Whose School Is It?": https://bilgimusterekleri.org/en/blog/egitimdosyasi/
• Knowledge Commons, "The Bell Does Not Ring the Same for Everyone": https://bilgimusterekleri.org/en/blog/okullaraciliyor/