It would leverage DeepMind's Frontier Safety Framework for procedural audits.
https://www.gstatic.com/marketing-cms/2f/c2/2f29fd1c44cfa0e7dd69fea9eec2/a-pragmatic-approach-to-ai-governance-in-america.pdf
Google just published an updated AI policy framework which articulates stronger and more detailed positions in some areas (e.g. auditing) than they have put forward in the past - glad to see!
No Digg Deeper questions have been answered for this story yet.
At Google we are proposing a frontier regulatory organisation (FARO) to promote frontier safety standards and oversee an independent audit regime.
The debate over AI governance is at a deadlock, with some calling for heavy regulation and others pushing for little or none to maximize innovation. We suggest a middle way designed specifically for the realities of both frontier models and widely-deployed AI applications. (1/2)
@HaydnBelfield Pharaoh, really?
At Google we are proposing a frontier regulatory organisation (FARO) to promote frontier safety standards and oversee an independent audit regime.