Sayash Kapoor testifies to California Assembly with layered AI safety legislative blueprint
sayashk · x · 2026-10-29
AI safety researcher Sayash Kapoor (incoming Berkeley professor) testified to the California Assembly Privacy Committee on AI legislation. Key observation: immense legislative enthusiasm — 15 assembly members, including the Speaker, attended an off-session event. His layered AI safety plan: strengthen liability (building on AB 316), require AI companies to hold liability insurance, mandate third-party safety audits, broaden transparency via incident reporting/safe harbors/whistleblower protections (below SB 53 thresholds), publicly support defensive AI in critical infrastructure, fund AI safety research (e.g., compute via CalCompute), and create coordination mechanisms so defenders get capability information from AI companies faster.
More from Safety
- Open models with Astra-level cyber skills: defenders-first access debate splits security researchers — rez0__ · 2026-10-29
- UW professors on the grand challenge of mitigating tech security, privacy and safety risks — lazowska · 2026-10-29
- Alpine Linux discloses GitLab security incident over attempted CVE-2026-97229 exploitation — sashk · 2026-10-29
- Sam Altman: models find hardware-level bugs 'outside the threat model' — 'we can't patch all of it' — DavidSKrueger · 2026-10-29
- Protecting action-taking agents from prompt injection: move defense out of the model layer — Informal-Dust4499 · 2026-10-29
- 26 academics publish working paper on feasibility of a hardwired pause on frontier AI training — _Dave__White_ · 2026-10-29