Academia: Self-Infliction

Thursday's Child Doesn't Have Far to Go If They Don't Cut it Out Pronto

I’ve always wondered what doctors think when they come to the conclusion that there’s a physician at another practice or hospital who is handing out bad diagnoses like candy, or is prone to small but significant errors in procedures and surgeries. Or more to the point, I wonder what they used to think and do. Did medical professionals effectively discipline themselves back when doctors were in charge, when nurses had authority over their own? I’d like to think so even though I know it wasn’t consistently the case.

Twentieth-century professionalization has a complicated history. Its motives and consequences were not straightforwardly positive across the board. You could argue that it was essentially a specialized mode of unionization, or even perhaps an anti-union kind of workplace organization that separated out professionals from mere “workers” and argued for a permanent wage differential favoring professionals, invoking a market logic to buttress that argument—essentially that the credentials that professional associations helped to define were in short supply both because of their difficulty—that not everyone was capable of acquiring the credentials—and because of the intrinsic costs incurred in the process of professional training and apprenticeship. But just as union labor often leads to far more skilled outcomes in trade services and factory production, the same was true for professional services in the second half of the twentieth century. Professionals who controlled their own workplaces and institutions also generally maintained high standards of service and care. People who turned to uncertified health care, accountancy, law, education and so on have often regretted it, much as people who turn to uncertified and uninsured craftworkers do—and the consequences when it’s your body, your money, your freedom and your education are more serious than a botched renovation or a bad rewiring.

For the same reason, the waves of deprofessionalization that have swept across the United States in the 21st Century have generally left almost everyone poorer, less healthy, less financially secure, less able to access legal services and less educated, often while paying much higher out-of-pocket prices for professional services of whatever quality. The only people better off are the people extracting value out of the systems that now employ professionals without allowing them much discretionary judgment about the services they deliver.

Under these conditions, if professionals ever did provide some form of quality assurance, many doubtless find it difficult now. When your expert judgment is routinely overridden or deflected by managers who have none of your training, why expect that the people in charge will welcome or listen to a report of malpractice? They’re just as likely to turn on the person making the report for exposing the institution to liability or regard expectations of ethical and dedicated professional labor as an unaffordable relic of days past. At least some of the time, a professional who is cutting corners or delivering poor service has been very nearly ordered to do so in the name of austerity, and the rest of the time, they’re responding to the relentless logic of incentive thinking, accurately sensing that there is no reward nor requirement for doing the job right.

And yet, and yet.

I do find it hard to watch those faculty across American higher education who seem eager to accelerate deprofessionalization, who don’t have some remnant of dedication to doing the job. Some faculty, after all, contributed to the rise of contingent labor in the previous generation. That shift didn’t all come from “the administration”, trustees or legislators. I grant that in its earliest stages, the connection between the move to adjunctification and various kinds of staffing decisions, curricular designs, and institutional practices might have been hard to see, and by the time higher education switched over, it may have felt equally hard to undo. (Though I think there always were—and still are—levers that people did not pull that might change the situation.

I’m far less forgiving about faculty who are now turning to LLM-based generative AI to write evaluations of student work, to construct slidedecks for lectures, to handle communications with students in a course, or to create professional writing of any other kind. There are other uses that I think are legitimate, if handled very carefully, but not any of that work. I understand that in times where faculty are frequently being treated badly by universities and colleges, the demand that professors ought to deliver hand-crafted, bespoke instruction that rests on their specific expertise might seem faintly absurd, and the idea that a professor like myself who works in extremely privileged conditions should stand in judgment of colleagues who have far more difficult jobs that are compensated far more poorly can seem distasteful.

Nevertheless, I’m begging professors who have done the kinds of things described in this week’s NYT article to cut it out. There’s so much wrong with this sort of practice. It destroys the last remaining shreds of legitimacy clinging to faculty demands of students that they not use generative AI, and contributes to an accelerating lack of faith in our profession at a time when we’re under a brutal attack from powerful enemies who want to destroy our institutions root and branch. It saves managers and administrators the effort of deprofessionalizing faculty: if you’re not using the expertise that your educational credentials say you have to prepare lectures, deliver work and evaluate performance, then it’s a short jump to hiring low-wage employees to simply read slides and from there to have the slides deliver themselves, to cut not just expertise but humanity out of the loop altogether. Buying a bunch of ready-made third-party content produced by big for-profit companies was already bad enough. We’ve been undercutting ourselves for decades with bad textbooks and worse packaged problem sets. Using information technology incompetently while having a lot of strong opinions about it has also been a big issue—for decades, many faculty have been complaining loudly about disinformation, misinformation, social media, and digital literacy without having any knowledge about the target of their complaints.

But this is worse. Much worse. It’s lower-quality material even if you can be bothered to clean up the hallucinations and the leftovers of the prompting process, which apparently many of the faculty who are turning to AI for these purposes can’t be.

Cutting corners with AI is just an elaborate way of cutting your own throat. If a lot of professors do it, just as professionals in other fields, they’re just collaborating in making it easier to get rid of expertise and knowledge work altogether. I understand why it’s tempting to give into nihilistic despair at this moment, but at least try to hold the line.