A guest lecture at NYU

On Friday, 23 October 2026, I gave a guest lecture at NYU's Washington Square campus, in Data Science for Social Impact, a course NYU Steinhardt requires of its applied statistics master's and PhD students. Professor Kate Schwartz, who teaches it, and Professor Jennifer Hill had come across my research and invited me to lead the class's session on AI, and Kate assigned Absent-Byte Diagnoses as reading beforehand. We started with a simple question: what would you let AI do on its own, with nobody checking? Then we looked at what happens when you ask a model about a chest X-ray you never actually sent, and at how the way a bias audit is set up can decide what it finds.

The rest of the session was about deploying AI in critical settings: the areas where we need to be extra careful with it, and the kinds of testing and checks that help it keep growing there safely. At the end we came back to the opening question to see if anyone's answer had changed. Thank you to Kate, Professor Hill and the students for having me. I had a great time.