Leiter Reports: A Philosophy Blog

News and views about philosophy, the academic profession, academic freedom, intellectual culture, and other topics. The world’s most popular philosophy blog, since 2003.

  1. Charles Bakker's avatar
  2. Anonymous's avatar
  3. UK academic's avatar

    In my department, we ran all postgraduate dissertations from last year through Pangram and found quite a few had generative…

  4. Charles Bakker's avatar
  5. Muhammad Ali Khalidi's avatar

    John Heil mentored a large number of philosophers by directing several NEH summer seminars on topics in philosophy of mind,…

  6. Randolph Clarke's avatar

    I studied with John when I was an undergraduate at Virginia Commonwealth University. Subsequently he was tremendously supportive of me…

  7. CT's avatar

    Update: Princeton appears to be GRE optional for this cycle. https://gradschool.princeton.edu/admission-onboarding/prepare/gre

Is AI devastating higher education?

A report from Australia:

As an academic, you encounter the process at the opposite end. When you open Speed Grader to confront another bland but competent assignment, suspiciously devoid of grammatical or spelling errors, your heart sinks. You know that, in all probability, the document you’re reading was composed not by your student (whose contribution likely extended no further than dropping the essay question into Claude’s interface and adding “make it sound human”), but by a legion of absent (and, quite probably, deceased) authors, whose forgotten texts enabled an algorithm to stitch together an essay that is at least passable and often reasonably good.

Grading induces a sense of utter futility: what could be less meaningful than offering feedback to dead people?….

One study of 8,000 people found 40% admitting they had cheated via algorithm. But that was in 2024 – aeons ago, given the rapidity of technological improvement.

A 2025 survey by the ARC Centre of Excellence for Automated Decision-Making and Society at RMIT identified students as particularly heavy users of AI. Nearly 80% employed the technology – though the study did not specify what for. Research on American undergraduates conducted in 2024 by UC Berkeley’s Center for Studies in Higher Education suggested two-thirds used AI. The following year, a UK analysis suggested generative AI was “near universal” among British students….

If you work at a university today, you talk a lot with the dead. I don’t mean that as a metaphor – or, at least, not entirely….

No one knows the proportion of Australian students reliant on AI for university assessments. One study of 8,000 people found 40% admitting they had cheated via algorithm. But that was in 2024 – aeons ago, given the rapidity of technological improvement.

A 2025 survey by the ARC Centre of Excellence for Automated Decision-Making and Society at RMIT identified students as particularly heavy users of AI. Nearly 80% employed the technology – though the study did not specify what for. Research on American undergraduates conducted in 2024 by UC Berkeley’s Center for Studies in Higher Education suggested two-thirds used AI. The following year, a UK analysis suggested generative AI was “near universal” among British students….

“Why memorise facts or do equations or slog it out over an essay when ChatGPT or Gemini can do it in seconds?” a Macquarie University student identified as “Hayden” asked rhetorically in a recent report by the Australian’s Ros Thomas:

You get everything in the blink of an eye – any equation, any answer, any essay. It’s all being hacked by AI and the sooner people accept that the better. Stop worrying about what’s right and wrong, because everyone’s using it and you’d be dumb not to.

Thomas presses Hayden as to whether he knows any students who complete their assignments without AI. “Nup,” he replies. “Not a single person.”

How are things where you are? Please identify the kind of institution (e.g., R1 university, SLAC, etc.) you teach at.

I have yet to read something written by AI by any of my students here (and I have checked with Pangram to confirm my impression).

,

Leave a Reply

Your email address will not be published. Required fields are marked *

One response to “Is AI devastating higher education?”

  1. In my department, we ran all postgraduate dissertations from last year through Pangram and found quite a few had generative AI signatures, ranging between 10% and 80% of the text. It’s difficult to know how seriously to take that, because text which is only very lightly edited (e.g., cleaning up the grammar and minor edits for readability) can get flagged. Personally, I wouldn’t care too much if the Pangram detection was only for light editing, but I suspect that in some cases (I have no idea how many) it was also for the argumentative content.

    I’m increasingly of the view that we’re going to have to drop all restrictions on AI use for dissertations. Why? Because it takes way too much time to enforce that policy and to run academic misconduct cases when we do find clear evidence to support such a claim. (It also exposes the academic tasked with prosecuting the academic misconduct case to allegations of bullying, by a student who follows the best-defense-is-a-good-offense strategy.)

    If we allow AI use on dissertation, the grade would then likely be based on an hour-long oral examination where we ask really challenging and probing questions about details of the argument, the literature cited, and responses to objections. That’s unfortunate, because oral examinations are biased against non-native speakers, some people who are neurodiverse, and other people who are just not natural public speakers, for whatever reason. But at least the oral examination would allow us to measure that there was some actual knowledge and understanding acquired at the end of the dissertation process, however it was acquired.

Designed with WordPress