Technology
OpenAI hacking incident exposes mounting risks in AI arms race
Key Points
Increasing use of aggressive training techniques sharpens threat of bad behaviour by leading models
Increasing use of aggressive training techniques sharpens threat of bad behaviour by leading models