Did We Just Award a PhD to an AI? Hard Calls in Doctoral Advising

About

50-minute Workshop

Presenters: Eric Freudenthal, Associate Professor, Computer Science; Natasha Rovelli, Doctoral Student, Computer Science

Did we just award a PhD to an AI? You can't reliably tell. Banning ChatGPT, Claude, or Gemini may just drive their use underground. And someone still has to sign the dissertation.

Our students will use these tools whether we permit them, like them, or detect them. So the real question is how we raise the odds that they still learn the analytic and communication skills a PhD is supposed to represent. Which skills must a graduate demonstrate without help? Is AI assistance really so different from the rewriting many advisors already do? And what should we spend our advising time coaching: the reasoning, the writing, or both?

Students face the other side of these decisions. Co-leader Natasha Rovelli, halfway through her doctorate, wants to become an effective scientist and build her career with dignity and honesty. Her question: with limited time, how should she divide it among the science itself, learning to structure a written argument, and finding and reading the literature? And where can she honestly get help from AI?

This is not a lecture. You'll work through realistic advising dilemmas with colleagues, see where the room splits, and leave with at least one practice to try. Bring your hardest case and your strongest opinion.

In keeping with its subject, this workshop was designed with the assistance of Claude, an AI model, and the facilitators are happy to discuss exactly how.