Meta plans to automate up to 90% of its privacy and integrity risk assessments, shifting from human reviewers to AI. This change, aimed at faster product launches, concerns current and former employees who fear reduced scrutiny of potential harms. While Meta claims human oversight remains for complex issues and that AI decisions are audited, internal documents suggest automation extends to sensitive areas like AI safety and youth risk. Critics argue that prioritizing speed over human review could increase risks, particularly given engineers' limited privacy expertise and the company's past record. The EU may be partly exempt due to its stricter regulations.

Prepared by Jonathan Pierce and reviewed by editorial team.