Policy2h ago
DeepMind CEO Proposes AI Model Vetting Body
Demis Hassabis is lobbying Washington to create a self-regulatory body for AI, mirroring the structure of FINRA. This plan emphasizes voluntary pre-release testing to identify risks before models reach the public. If adopted, DeepMind and its peers would shift from internal audits to a standardized, industry-wide vetting process for safety compliance.