Provably Safe AI - Steve Omohundro

Provable AI Safety, Steve Omohundro - May 2024
​We'll discuss an approach to AI safety based on the laws of physics and mathematical proof which are the only two guaranteed principles for constraining powerful AGI systems. The approach is very promising but will require many new ideas. We'll discuss how the technology could not only lead to AI safety but also to human thriving.
Slides: drive.google.c...
---
About the series:
​​​​The monthly seminar series on Guaranteed Safe AI brings together researchers to advance the field of building AI with high-assurance quantitative safety guarantees.
www.horizoneve...