That commercial AI bots like ChatGPT are the scourge of higher education is a well-rehearsed subject. If cheating was rampant before this technology came on the scene — in the years prior to my retirement in 2023, I already was confronted with significant numbers of students copying things off of Wikipedia or subject-specific online encyclopedias — it has gone through the proverbial roof now. Professors regularly report that they face a deluge of AI generated papers of ever increasing sophistication, which makes detection far more difficult than it once was. I used to be able to catch cheaters by copying and pasting suspicious text into a search engine, but this form of academic dishonesty requires a person to employ advanced software in order to uncover it, meaning that teachers have to spend more and more of their time functioning as digital sleuths. At their wits end, some faculty have returned to the good old Blue Book and in-class written exams, while others have opted to assess their students orally, though this remedy is only available to those in the swankiest of positions, enjoying the cushiest course loads. (Such a thing would have been impossible for me in my small, undergraduate-only department, where I regularly had sections of 90 students and taught over 200 per semester.)

While all of this is terribly concerning, because it matters whether students actually learn what they are purporting to study (imagine having an accountant who did all of his assignments using ChatGPT and never actually learned accounting), things have taken an equally dark turn on the production side of academia. What I’m talking about is professors using AI in conducting their research: more specifically, humanities professors and in particular, philosophers. If you’d told me five years ago that academic philosophers would employ chatbots in conceiving and writing their papers, I wouldn’t have believed you, but if you’d also told me that they’d be proud of it and that professors would twist themselves into mental pretzels defending it, I’d have thought you an overly negative sort or wondered if you might be insane. Yet, this is exactly what is happening today, and it will be the ruin of academic philosophy if it isn’t stopped soon.

Enter Jimmy Alfonso Licon, a faculty member in the philosophy department at Arizona State University which, it should be mentioned, is not some inconsequential service department out in the sticks, but a PhD-granting program at a major institution. He just published an essay over at Justin Weinberg’s Daily Nous (a much-read philosophy insiders blog) on his use of ChatGPT in his research, with the title: “Have Pen, Laptop, and ChatGPT, Will Publish” (a takeoff on Robert Heinlein’s 1958 juvenile novel, Have Spacesuit–Will Travel), so I’ve neither “outed” the man nor broken or betrayed some confidence. He is broadcasting what he is doing and has therefore invited whatever scrutiny it may receive.

The title involves a not very subtle conceit, the aim of which is to give the impression that using a chatbot in writing one’s academic papers is just like using a pen or a laptop. He repeats the formula in the very first paragraph, after which he and his defenders in the comments go on to make more exotic (and even less plausible) comparisons, suggesting that using LLMs in one’s academic work is no different from employing students as research assistants or brainstorming with one’s colleague down the hall. At one particularly heady moment in the back-and-forth, Nicholas Delon of the College of Charleston begs everyone to “let a thousand flowers bloom,” observing that “Lucretius wrote in verse, Nietzsche in aphorisms” and that “younger scholars use widgets,” finally wrapping up this collection of tangents and non sequiturs in a glorious, discursive ribbon: “Let people experiment and judge the results.” It’s all very weird stuff not to mention quite disconcerting, insofar as it comes from the mouths (or in this case, the keyboards) of people who are supposed to be deep and critical thinkers.

Licon’s account of the ways he uses ChatGPT in research puts the lie to the idea that it’s no different from using a pen or a laptop. I counted at least five points in the writing process in which he consults the chatbot, and it is quite clear that for him the software is anything but a mere tool. Indeed, Licon essentially admits as much at the end, concluding that “the resulting workflow is neither purely human nor AI-written,” which undercuts most of the apologetics in the discussion that follows and makes the raison d’etre for the entire exercise unobvious, but never mind. Those of us who are desperately worried about the idea that philosophy professors have joined the slide of our population and institutions into the aggregating, homogenizing, brainless maw of AI are just engaging in “whiny, irrelevant, cope,” as he puts it in a moment of pique during the discussion following the piece.

Licon’s research apparently begins with a “nagging thought” or a “question that won’t leave me alone” (❡ 2) that subsequently inspires an idea, which he then shapes into an abstract. Then, he has ChatGPT [1] “analyze the idea like a referee at a top journal,” and [2] “hunt down the relevant literature” (❡3), which is how he determines whether he has something “new, defensive [sic], and interesting to say.” (❡4)

The next thing Licon does is read some of the literature the bot has suggested (as well as Google Scholar) and endeavor to “capture what will matter for my argument.” (❡5) After doing this “cognitive and analog [?] legwork,” he once again employs ChatGPT [3] in the mode of an “overcaffeinated colleague who has read a lot, reasons quickly, and is perfectly happy to brainstorm for as long as I like, but whose judgment I do not automatically trust.” (❡6) 

He returns to the chatbot once again, this time [4] to “write a rough synthesis of the project on the basis of the abstract and my summaries of the research.” (❡8) The point of this, he says, is to “see my own materials in a different light to highlight which parts of the project are doing real work and which are just taking up space because I happen to find them interesting.” (❡8)

After writing a draft, Licon consults ChatGPT one last time, albeit in a “very different role.” He tells us that [5] “at this stage, I ask it to behave not like a chatty colleague but like a demanding referee. I paste in the draft and explicitly request a rigorous, critical report…I encourage it to be uncharitable, and to read the paper like an overworked, skeptical reviewer might.” (❡12)

There is much here to alarm and dismay, and several critics (of which, sadly, there are few) express some of it in the comments. By now, it’s well-understood that LLM’s like ChatGPT are trained by way of the widespread theft of intellectual property, which means that when one uses it in one’s own writing, one is engaged in a kind of indirect stealing, several steps removed. (This is just one of the areas in which there is a clear disanalogy between using ChatGPT and talking to one’s colleagues, the latter of whom presumably consent to having their brains picked.) There is the further matter of what is sometimes referred to as “cognitive atrophy,” meaning that the more our mental capacities are outsourced to a chatbot, the weaker they will become over time, a concern that is lampooned by Licon’s defenders, like Kenny Easwaran of UC Irvine (another ostensibly serious place) who quips that “everytime you ask a friend for help…you’re letting some of your skills atrophy” and claims that what he is concerned about is that “people are too willing to let their spatial navigation skills atrophy and just offload all navigation to a GPS system,” which one hopes is just an inept effort at sarcasm and not intended to make any sort of serious point.

But what I find most upsetting about the whole thing — both the essay and the discussion afterwards — is just how unrecognizable the entire enterprise of philosophy has become, and not because professors are using electronic tools rather than poring through the pages of The Philosophers’ Index or spending hours or days in the library stacks, as my generation did.

For one thing, it is clear that the author and his defenders think the social dimension of writing philosophy is entirely dispensable; that the discipline descended from Socratic interlocution is now something that one does by oneself with a machine. How else to understand the impetus to substitute a chatbot for interaction with one’s colleagues, when one is in a department with twenty-five other scholars? (In the thirty years that I spent in my tiny department, which fluctuated between having five to seven members, I used to fantasize about being in such a large program with so many peers to interact with.) The teaching and mentorship dimension of academic research is also cavalierly tossed aside, as evident from the bizarre (and tone-deaf) comparison between consulting a chatbot and employing graduate or undergraduate research assistants. As a graduate student at the CUNY Graduate Center in the 1990s, I benefitted enormously from my research assistant gigs with Hartry Field and Steven Cahn, and several of my undergraduates at Missouri State got a great deal out of research assistantships with me, which I effected through the creative use of the “Independent Study” course designation, since we had no money to pay them. [Since publishing this essay, Eric Steinhart of William Patterson University has deemed research assistantships “stupid grunt work,” demonstrating that beyond sometimes being foolish and obtuse, academic philosophers may also be rather unpleasant people.]

For another thing and whatever the reasons (I have suspicions, but will leave them for another occasion), the view that philosophy is primarily a technical, problem-solving enterprise seems to have taken over the discipline. Only a person in the grip of such a view could suggest, as Licon and his apologists do, that all that matters is “the results” regardless of how one arrives at them. That this is not what philosophy is becomes obvious if one spends even two or three minutes meditating on the sorts of questions we take up and the answers that have been offered over the millennia. I’ve written about this extensively already, so I won’t repeat myself much here, but if you think that the questions of moral philosophy, political philosophy, action theory, epistemology, metaphysics, and the like are going to receive conclusive answers that command any sort of consensus among experts, then you haven’t been paying attention. Worse, you’ve lost the plot entirely.

Philosophy is much more akin to literature than it is like chemistry or engineering, in that it is as much an expression of the philosopher’s distinctive personality, temperament, outlook, and voice as it is anything to do with the objective features of whatever topic is at issue. When we read philosophy, we are invited into the writer’s mind; into his or her point of view; into a distinctive way of seeing and conceiving of the things that matter to us the most. The greatest philosophers have proven compelling in this way to a broad range of readers and thinkers, over vast stretches of time. Lesser philosophers less so. But, this dimension of what we do is not only not dispensable, it is essential. And once we have purged it through the use of this wretched technology, we are going to discover that the Academy, not to mention the broader public, will have little to no use for us and rightly so.