Paul Christiano: Staying in Control of Self-Improving AI | The Hinton Lectures 2026 (Night 2)
The Hinton Lectures YouTube 18,374 views as of 13 November 2026
Why it is here
The second 2026 Hinton Lecture (1 h 27 min), introduced by Geoffrey Hinton: Paul Christiano on whether human oversight can keep pace once AI helps build the next generation of AI. Long recording: per the owner’s rule (videos over ~1 h) it is not sent to Gemini; this note is written from the official description only (no transcript was available to us).
Description
Per the channel’s description: the lecture covers how AI could accelerate its own development, growing reliance on these systems and the limits of existing safeguards, and asks what it would take to keep humans informed and accountable. Lecture from 11:00; panel from 57:14 with Christiano, Jodie Wallis (Global Chief AI Officer, Manulife) and Jesslyn Dymond (TELUS), moderated by Sínead Bovell. Live from the John W. H. Bassett Theatre, Toronto, Nov 10, 2026.
Related
More videos on this
-
Paul Christiano: AI Alignment - What’s Working and What’s Not | The Hinton Lectures 2026 (Night 1)
The Hinton Lectures40,943 views as of 13 Nov
-
Paul Christiano: Governing Transformative AI | The Hinton Lectures 2026 (Night 3)
The Hinton Lectures12,242 views as of 13 Nov