AI safety via debate

8 years ago 27
Add to circle
We’re proposing an AI safety technique which trains agents to debate topics with one another, using a human to judge who wins.
Read Entire Article
  1. Homepage
  2. Technology
  3. AI safety via debate

Related

Anthropic’s New Agent Has No Cute Mascot, and You Probably Can’t Try It Yet

Anthropic’s New Agent Has No Cute Mascot, and You Probably C...

6 minutes ago 1
AuxoAI, which helps large businesses build and deploy AI systems, raised a $20.7M Series A led by NTTVC to accelerate growth and scale (Kyt Dotson/SiliconANGLE)

AuxoAI, which helps large businesses build and deploy AI sys...

11 minutes ago 1
YouTube says it isn’t testing hiding the dislike button

YouTube says it isn’t testing hiding the dislike button

14 minutes ago 1