When Mind and Phil Review are publishing papers by Claude 7.3 and GPT-14.5

In some ways, the Age of AI will make the humanities more important than ever. Disciplines from literature to philosophy are needed to help us answer questions about how we can find a place in the world when we are much less needed than before, and what it is to be human when we are no longer the only ones capable of doing some of the things of which our species was once uniquely capable. But at a time when artificial intelligence can jump through the hoops that have over the past decades come to define an academic career in the humanities with growing ease, a radical reimagination of how we pursue and impart meaningful knowledge in these fields is desperately in order.—Yascha Mounk

One of the most irritating features of my home profession(s) — philosophy and political theory — is that prestige-production is regulated by maintaining incredibly low acceptance rates to the most important academic journals; these are below the rates common in many other academic disciplines. Such prestige production matters as a (defeasible) “signal” in order to obtain a position and promotion at universities that care about prestige (and — not trivially — often have excellent jobs).¹

Loyal readers know that I lost my battle against my profession’s practice. (My best piece on the topic is probably this one (here)—but I have been railing against it throughout my blogging life.) This despite the fact that these very, very low rates of acceptance slow down the advancement of knowledge, generate huge opportunity costs to the field, including people who have no interest or hope in obtaining a position at the most prestigious departments, and they give a very narrow group of gatekeepers — usually a senior scholar who is considered a ‘safe pair of hands’ — concentrated power to make and break what counts as great scholarship.

Anyway, I am not here going to relitigate all that. So why return to the topic?

Yesterday, Yascha Mounk published, “The Humanities Are About to Be Automated: AI can now write convincing academic papers. There’s no more room for denial.” Since I also like long titles, the piece immediately endeared itself to me. The main point of his piece is a kind of existence proof: commercially available LLMs are (on the cusp of being) capable of writing journal article in the humanities with relatively little guidance from a human. (Yascha had not heard the news yet that titles with ‘Epistemic’ in them have been banned at the start of the year, but that’s a minor blemish.)

As he puts it this will require “a radical reimagination of how we pursue and impart meaningful knowledge in these fields.”² The second, sly polemical point is that “Professors at top universities are—in good part because of what it now takes to get a good job in the field—more focused on publishing erudite contributions to niche debates in scholarly journals which only a handful of their colleagues will ever read than on teaching and mentoring the ever-dwindling ranks of their students.”

As his polemical point hints at, Glaude 6.5, Gemini 4.3, GPT 13.6 are about to sweep away the old order just as undergraduate enrollment and funding for PhD students is collapsing in many humanities Stateside. (Well not all Humanities; philosophy undergraduate enrollments have stabilized during the last decade.) Let’s keep that in mind in what follows. I will address it more at length some other time.

What to make of this?

So, first, unsurprisingly given my entrenched views about our journal system, I would welcome the effective teasing apart of prestige production that functions as a signal for top jobs and the publication and circulation of knowledge.³ It can be done. In the past, there were different ways of ensuring that scholars exhibited prestige-inducing character traits (skill at calligraphy, ability to speak a dead language fluently, etc.) that were only tenuously connected to the advancement of knowledge itself.

I suspect that some Humanities fields are already experimenting with valorizing work that are outside the foreseeable margins of the capabilities of LLMs. Say, scholarship on objects that have no pre-existing secondary literature (I use that trick in my undergraduate honors course), or on ephemera, or that valorize means of expression that are themselves interpretive or idiosyncratic in character. Maybe the list is longer?

Second, I will be sad if it becomes impossible to receive recognition from one’s scholarly peers for one’s scholarship when nobody will feel confident that one’s scholarship is truly one’s own. Given my vanity, the demise of the existing credit economy in the Humanities will be rough going for me.

But this loss may well be outweighed by the joy in the growth of knowledge and means to satisfy one’s inquisitiveness. It may well be a good thing, too, that one’s ego is removed from the practice of scholarship which may return to its spiritual roots. I may come to have a disinterested pleasure in reading Claude 8.3’s monographs (with fancy graphics).

Third, I will not mourn the eventual demise of the cartel of for-profit academic publishers. If academic publication becomes devalued as job-currency, their business-model will not have a long duration. Undoubtedly, they will try to hang on in an arms race with AI publishers, trying to certify publication in their journals as LLM free. But that’s not going to last long, if only because Alphabet Corp (or whatever it’s called) will not wish to share profits.

Let’s return to Mounk’s essay. There are two linked questions lurking here: can “top universities” do without the “prestige” associated with the Humanities? And can “top universities” do without “prestige” in distributing jobs in the Humanities? I am genuinely unsure about the former question, although there are grounds for concern if the answer is ‘yes’ because we may see the elimination of the humanities from such universities (recall those dwindling numbers). Mounk is actually more optimistic because he thinks the social demand for Humanities education will rise in an age of AI. I hope he is right.

On the latter question, I am also unsure; but I doubt that teaching and mentoring is the right sort of activity to generate prestige. What follows is not intended to cast aspersion on teaching and mentoring. (I also happen to think teaching and mentoring are joy-inducing in teachers and mentors.) After all, what’s needed for the production of prestige is a mechanism that generates scarcity in addition to requiring non-trivial resources (time, money, network, connections) to succeed. If prestige is skill-based, one also needs to be able to discern fine-grained distinctions in the product. Most of these conditions are not present in the skill of teaching and mentoring, which does not require the kind of substantial endowment effects which I have hinted are needed for the production of prestige. That may be unjust, but that’s a feature not a bug of even the most ‘meritocratic’ prestige-systems.

Of course, at the top they may revere journal publication for a while longer. To paraphrase an old saying, it won’t be the first time that worship lags a changing of the guardians.

1. As Anna Stilz noted during one of the controversies, publication in these journals is a “signalling device that validates the importance of someone’s work when they go up for tenure and promotion.” [HT Dailynous]

2. It is worth noting that Mounk’s use of ‘radical’ here is echoed by the LLM in the abstract (and body of the argument) of the paper. This is surprising stylistic tic they have in common.

3. Arguably, publication in law schools already does this.