AI lab CEOs including OpenAI, Google, and SpaceXAI have rallied behind Anthropic's proposal for third-party audits of AI safety practices, but security expert Katie Moussouris argues the focus should instead be on foundational network security like logs and permissions. She compares the audit proposal to sidestepping the kind of 'Trustworthy Computing' mindset Microsoft adopted in 2002, suggesting basic security hygiene may be more impactful than external verification.
- Source
- TechCrunch
- Published
- Sep 17, 2026 at 02:25 AM
- Score
- 6.0 / 10