From Computerphile.

As AI systems become more capable, rule-based safeguards, hard-coded restrictions, and simple alignment strategies start to break down. Buck Shlegeris talks about some tactics we might use as detailed in a recent paper.

The referenced paper: https://arxiv.org/abs/2504.10374

Computerphile is supported by Jane Street. Learn more about them (and exciting career opportunities) at: https://jane-st.co/computerphile

This video was filmed and edited by Sean Riley.

Computerphile is a sister project to Brady Haran’s Numberphile. More at https://www.bradyharanblog.com