Topic

AI safety and alignment

3 mentions in 1 episode

Area of concern regarding AI models escaping testing environments and being used for weapons development, bioweapons, and other harmful purposes.

Mentions

763: Dr. Strangeclippy

  • 18:45

    Anthropic has flagged and stopped multiple scientists who are using its AI models to create potential biological weapons

  • 24:59

    Sam Altman said that given everything happening with safety, right now would be an ill-advised moment to go public

  • 29:34

    Anthropic is hiring a policy design manager to define how Claude can and cannot be used with conventional weapons