Slate essay warns colleges against ChatGPT surveillance fix

The argument worth pulling out of this Slate Future Tense essay by Chris Gilliard and Pete Rorabaugh is not really about ChatGPT. It is about the reflex colleges reach for when a new technology scares them, which is to buy more surveillance.

The setup is familiar. The Atlantic ran pieces headlined 'The End of High-School English' and 'The College Essay Is Dead.' A research paper made the rounds claiming GPT had passed a Wharton MBA exam. Into that panic stepped the obvious commercial answer, AI to detect AI. A Princeton student, the authors note, 'spent a chunk of his winter break creating GPTZero,' the app pitched as a way to tell human writing from ChatGPT output. Turnitin, already embedded in most campuses, is rolling its own detection in. OpenAI itself has floated watermarking and a classifier, while conceding the tool is not fully reliable.

Gilliard and Rorabaugh's pushback is that this is the wrong fight. They describe the underlying mood as a 'Borg complex,' the assumption, as they put it, that 'resistance to a particular technology is futile.' Once you accept that framing, the only question left is which vendor to pay, and the answer ends up being whoever is most willing to invade, in their words, students' 'kitchens, living rooms, and bedrooms.' Their alternative is pedagogical rather than technical. Writing, they argue, is a 'metacognitive process' of having an idea, composing it, and checking it against what you meant, and 'to outsource idea generation to an A.I. machine is to miss the constant revision that reflection causes.'

The honest caveat is that this is an op-ed, not a study. The authors do not give you measured accuracy figures for GPTZero or Turnitin, they do not name universities that have signed detection contracts, and they do not say what their reflective alternative looks like in a large lecture course. What the reporting does not give you is the operational picture, only the framing one.

The forward-looking piece is the regulatory analogy they reach for at the end, that society 'took the important step of regulation in other major industries (like tobacco, pharmaceuticals, automobile manufacturing).' Whether that frame catches on with administrators, or whether the detection vendors win the procurement cycle first, is the part to watch.