Three closures, one rescue and a haemorrhage that nobody has measured
Dundee has suspended its single honours philosophy degree and approved a philosophy-and-AI programme instead; Hertfordshire has stopped recruiting; Kingston's research centre has left the university altogether. Meanwhile the labs hire, Floridi speaks of a 'haemorrhage' and the Fed's table is quoted as vindication. An audit of what the year has actually established.
Tower Building, University of Dundee · Wikimedia Commons
On 27 August the University of Dundee's press office published a news item under the headline "From Dundee to Claude's constitution: philosophy and AI". It celebrated Amanda Askell, who studied philosophy and fine art at Dundee and who was the principal author of the constitution that Anthropic published in January to govern the behaviour of its Claude models. Nine weeks earlier the same university had proposed to stop offering philosophy as a named degree at all. The rescue, when it came on 5 September, preserved the joint degree in art and philosophy, which happens to be the one Askell took, and suspended the single honours degree, which is the one in which philosophy is studied on its own account.
That sequence is the story of the year in miniature, and it has been told, in this publication and elsewhere, as a paradox: the laboratories hire philosophers while the universities shut philosophy down. Our earlier reports on the closures at Dundee and Hertfordshire and on the hiring by Anthropic, DeepMind and the rest both drew on the news pages of Philosophy Now's August issue, Three months on, the framing has acquired a stock quotation from Luciano Floridi and a stock statistic from the Federal Reserve Bank of New York, and it deserves an audit: what was actually decided, by whom, on what date, and how much of the surrounding argument survives contact with the record.
The claim in circulation is that four British universities have closed or cut philosophy within a year. The count does not hold.
Kingston University's decision is the oldest. On 6 March 2025 Times Higher Education reported that Kingston would close its entire humanities department, admit no further students to English or philosophy, and with them close the Centre for Research in Modern European Philosophy. The university cited rising costs, falling overseas applications and a competitive market, and added, with unusual candour, that the affected programmes "have had to be subsidised by others across Kingston School of Art for a number of years". The Kingston branch of the UCU called the thirty-day consultation "unreasonable and unlawful" in an open letter of 19 March 2025, with about thirty staff at risk. None of this happened in 2026. CRMEP had already been in trouble in February 2024, when Kingston suspended admissions to its doctoral programme and a letter signed by Jürgen Habermas, Judith Butler and Slavoj Žižek described the centre as "one of the most valuable and distinctive centres of philosophical research anywhere in the English-speaking world".
The German Society for Philosophy has launched a working group on philosophy and National Socialism, conceding that the subject has no settled place in research or teaching. In Beijing, sixty scholars debated how to build a 'Chinese autonomous philosophical knowledge system', a phrase that belongs to Xi Jinping. And a sociologist's book on reproduction in Palestine/Israel reaches a philosophy audience.
What has happened since is the development worth reporting. CRMEP's website now describes it as "an independent educational association based in London" that organises "events, publications and study groups", and it has announced a public lecture series at King's Cross, on Thursday evenings from 15 October, with Éric Alliez, Howard Caygill, Peter Hallward, Catherine Malabou, Peter Osborne and Stella Sandford. Kingston School of Art's own page of research centres no longer lists it. The centre was founded at Middlesex in 1994 by its own account, moved to Kingston in 2010 after a student occupation and an international outcry, and has now left the university system altogether; we could not establish the date or terms of its departure. A research centre with no undergraduate income and a heavy doctoral output has become a voluntary association with a lecture programme. It is a different form, not obviously a worse one, and nobody chose it.
Hertfordshire is a 2026 decision. On 6 May Times Higher Education reported that the university had suspended recruitment to history, philosophy, English literature, linguistics and creative writing, citing "sustained low student recruitment and declining demand". A spokesperson called these "difficult but necessary and responsible decisions" and said the focus remained "on supporting current students to complete their studies", which is the language of a teach-out. The local UCU branch said that "humanities colleagues have not seen any of the metrics used for the decision process". Nothing was formally closed. Recruitment was stopped, which produces closure by attrition three years later, without the decision ever having to be defended as a closure.
Dundee is the case with the fullest record. The Strategy to Recovery proposals were published on 18 June: philosophy would cease to be offered as a named degree from 2027, philosophy content would be retained "as part of a broad MA undergraduate programme, at least until 2030", and a forty-five-day collective consultation would run alongside twenty-one full-time-equivalent job losses in the Faculty of Arts and Social Sciences. The university's stated reason was "very low numbers". On the same day Oisín Keohane, a lecturer in the department, opened a petition on Change.org. The Courier had it at nearly 3,000 signatures on 26 June; The Scotsman at more than 5,300 on 15 August, including over 1,100 academics; the Change.org page showed 6,130 verified signatures this week. The signatories included Floridi, Malabou, Žižek, Shannon Vallor, Philip Pettit and the broadcaster John Suchet, who read philosophy at Dundee in the 1960s and offered the campaign its best line: "Dundee University will be less of a university if it ceases to offer philosophy as a degree."
The outcome, reported by The Scotsman on 5 September, is more particular than "saved". Single honours philosophy is suspended, and so are the joint degrees with politics and with languages. Art and philosophy, English and philosophy, and philosophy and psychology continue. Undergraduate and postgraduate qualifications in philosophy and artificial intelligence have been approved for development. The department's message to supporters said "the ground shifted because of you" and that "a small programme with a large international reach has bucked the trend", and then that "the picture is not all good", because the decision had been reached without "clear financial rationale or assessment of the risks". The university still plans 154 redundancies. What the detail shows is that philosophy survives at Dundee wherever it is attached to something else, and nowhere on its own.
So the honest count is not four closures in a year. It is one closure completed in 2025 whose research centre has since gone independent; one suspension of recruitment in May 2026 that will become a closure by 2029; and one proposal in June 2026 that was partly reversed in September. Three institutions, three mechanisms, three different remedies available.
It is easy to write about these decisions as philistinism, and mostly wrong. Kingston's admission that its humanities programmes had been cross-subsidised for years is a fact about English funding that every registrar knows and few say aloud: a classroom subject with no high-cost teaching grant, whose income is headcount multiplied by a fee frozen in cash terms for most of a decade. Our analysis earlier this month of what the labs are buying and what the universities are shutting set out that arithmetic in full, and nothing in the three cases contradicts it. Dundee said "very low numbers". Hertfordshire said "sustained low recruitment". Kingston said cross-subsidy. Nobody has produced enrolment figures to show any of them was lying.
What the campaigners can fairly say is something narrower and more damaging: that the numbers were never shown. Hertfordshire's staff had not seen the metrics. Dundee's philosophers, having won, complained that the reversal came with no financial rationale either, which suggests that the decisions in both directions were made on something other than a published subject-level account. That is the strongest version of the case against the vice-chancellors, and it is a case about process rather than about the value of Hegel.
Where the word "haemorrhage" came from, and what its author now says it means#
The Floridi quotation has become detached from its source, so here is the source. The Economist published "Why big AI labs are hiring so many philosophers" on 24 June, and quoted Floridi, who holds a chair at Yale, describing the scale of departures from philosophy departments as a "haemorrhaging". Brian Leiter responded the next day on his blog, calling the claim "obviously nonsense: there are still far more philosophy PhDs than there are jobs, even if a handful are leaving academia for the AI world". Floridi replied in his own way on 6 July, in a Medium essay in his "notes to myself" series, which begins by acknowledging that "I have been quoted describing the departures from philosophy departments as 'haemorrhaging'" and then does something more interesting than defend the word.
His argument is that the laboratories are not buying expertise in ethics at all. They are buying a temperament: the capacity to work inside a dilemma without forcing it closed.
You cannot optimise your way out of a dilemma. You need people trained to live inside one.
And the cost to departments, on his account, is structural rather than numerical. "Success bleeds talent; failure bleeds students. So the better the discipline becomes at its work, the more it stands to lose by it." Leiter is right that a few dozen departures cannot drain a labour market that has been in permanent oversupply since the 1970s. Floridi's reply, in effect, is that the departures are not random: the people who leave are those the departments can least spare, the ones capable of the applied work. Whether that is true is an empirical question about who exactly has gone, and nobody has answered it.
Nobody has the numbers at all. Benjamin Sutherland, who wrote the Economist piece, told NPR on 7 July: "I wasn't actually able to get good numbers on how many are being hired, but I think they actually are in pretty senior positions." The named cases are real. Askell leads the team at Anthropic responsible for Claude's character and wrote its constitution. Keohane, in Times Higher Education on 17 July, wrote that "DeepMind and Anthropic between them now employ numerous philosophers doing philosophy as their actual job". OpenAI's contribution is a document rather than a named hire: its Model Spec, revised in August, is a work of applied casuistry drafted by a committee on which philosophy is one voice. A handful of senior people at three firms, plus a drafting committee, is a fact about three firms. It is not a labour market, and it is not a reason for a seventeen-year-old in Hatfield to choose the subject.
The statistic that travels with the Floridi quotation comes from the New York Fed's table of labour market outcomes by major, which is updated each February and defines early-career graduates as those aged 22 to 27. In the February 2026 release philosophy's unemployment rate was 5.1 per cent and computer science's 7.0, with underemployment, meaning work in a job that does not normally require a degree, at 47.1 per cent for philosophy and 19 per cent for computer science. Median early-career pay was 52,000 dollars against 87,000. Those are the numbers, and the underemployment column is the one to read first.
The detail that has gone unremarked is what happened between releases. When the table made headlines in May 2025, from the previous year's data, philosophy's unemployment rate was 3.2 per cent and computer science's 6.1, and the story ran as "philosophy now outranks finance". A year later philosophy had risen to 5.1, and the story ran as a triumph because computer science had risen further. The philosophy figure got worse and was celebrated for it. Given the sample sizes, the honest reading of both vintages is that they are employed at about the same rate as everyone else and paid a good deal less.
Strip away the paradox and three findings remain, none of them paradoxical.
The first is that philosophy in Britain now survives institutionally as an adjective. Dundee keeps philosophy-and-art, philosophy-and-English, philosophy-and-psychology and will build philosophy-and-AI; it loses philosophy. Hertfordshire loses everything. Kingston's centre lives on outside the university. The subject persists where it is attached to a discipline with better recruitment or a word with better press, and the single honours degree is the form being suspended.
The second is that the "AI needs philosophers" argument has now been tested as a rescue instrument, once, with partial success. It worked at Dundee because the department could name a graduate, because Floridi and Vallor were willing to sign, and because the university could be offered a programme with AI in the title. It did not save single honours. Keohane's formulation in the Dundee press release is the most precise statement yet of what the laboratories think they are getting: "Philosophy supplies the content whilst the training loop supplies what philosophy has never had with people: a way of checking whether it took, and of trying again if it didn't."
The third is that the haemorrhage, if there is one, runs in the opposite direction from the closures. Kingston's centre and Dundee's department were two of the most visibly continental philosophy programmes in Britain. The philosophers the laboratories name are, for the most part, trained in the analytic tradition: decision theory, philosophy of mind, formal ethics. Whatever is being drained from departments into Anthropic and DeepMind, it is not what Kingston and Dundee were teaching. The closures and the hiring are two separate events that happen to concern the same word, and the sooner the defence of the subject stops relying on the connection between them, the sooner it will have to make the harder argument: that a university with no philosophy in it is, as Suchet put it, less of a university, and that this is worth paying to avoid. That argument has no Fed table behind it. It has only the case on its merits, which is where philosophy has always had to start.
Falling fertility and ageing in rich societies, and migration across every border, force an old question back open: who makes up a political community, and who gets to decide? A reading through Arendt, Carens, Walzer, Miller, Pogge and Parfit.
Will Kymlicka, Thomas Pogge, Philippe Van Parijs, Hannah Arendt
This afternoon in New York, under the French presidency, the Security Council holds its first meeting devoted to the possibility that frontier AI systems slip out of human control. The briefers are Yoshua Bengio and the chief executives of OpenAI, Anthropic and Hugging Face. Whether the Council is the right room is itself in dispute.
.jpg?width=500 "Kant")
Comments
0Loading comments…