Discussion

AI is challenging what universities measure, not just how students write

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#15028

Generative AI is forcing universities to confront an awkward question: if students can produce coursework with a prompt, what are grades and degrees actually measuring? In an UnHerd essay, Richard Hanania argues that AI may expose how much higher education relies on credentials and routine assessments, rather than simply destroying learning.

Watch Desk analysis

What happened

Hanania says AI can produce passable essays and coursework quickly, but can also act as a tutor. He points to a review of more than 75 studies by David Deming, an economist and Harvard College dean, which he says found that AI generally helps learning when used in structured ways, such as tutoring or feedback. Hanania’s piece is an argument about what those findings mean, not a neutral survey of higher education.

He suggests that universities could respond with more proctored exams and in-class essays, or adapt to AI-assisted work. The essay also cites a study of an Israeli university in which 62% of professors said they had changed assessments in response to AI, while researchers found almost no change in course syllabi in the first two years after ChatGPT’s release.

Why it matters

The practical question is bigger than whether students use AI to write. If take-home assignments no longer show what a student can do independently, institutions may need new ways to assess learning. But treating every AI-assisted answer as cheating could also shut out tools that, as the studies Hanania cites suggest, can support learning when used deliberately.

That leaves universities juggling two goals: helping students learn and deciding what a qualification should certify. AI has made that tension harder to ignore, though it has not settled Hanania’s broader claim that higher education is mainly a signalling system.

Our read

The useful part of this argument is its challenge to assessment, not its sweeping verdict on the value of university. A degree can signal persistence and still involve real learning; the difficult job is working out how to demonstrate both. Universities should be judged by whether their assessments reward understanding, not by how quickly they can ban the latest tool.

What to watch

  • Whether universities replace take-home work with more supervised assessment.
  • Whether AI tutors improve learning in studies that measure outcomes, rather than just use.
  • Whether employers and universities change what they expect a degree to prove.

Discussion spark: Should universities redesign assessment around what students can do with AI, or preserve exams that test what they can do without it?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.