Bipartisan US lawmakers outline SAFE AI Act: mandatory safety frameworks and incident reporting
ShakeelHashim · x · 2026-11-20
Reps. Beyer, Jacobs, and Maloy released an outline for sweeping bipartisan AI safety legislation, the "SAFE AI Act" (details via screenshots):
- Mandatory safety frameworks and disclosures: frontier developers must publish enforceable safety frameworks filed with CAISI, disclose substantially autonomous AI R&D, and report safety incidents and imminent severe-harm risks; anyone can confidentially report incidents to CAISI.
- Truly independent evaluations with government oversight: government teams examine each developer's frameworks; accredited third-party verifiers conduct compliance evals with fee-based funding caps, mandatory rotations, and pre-deployment evaluations covering commercial release and automated AI R&D.
- Enforcement: narrow emergency authority with judicial review; civil penalties and criminal referral procedures; clarified developer liability for foreseeable third-party injuries; whistleblower protection.
Related event: US Bipartisan Lawmakers Unveil SAFE AI Act Framework(3 posts)→
More from Safety
- UK court publishes full judgment in Apple Inc v Home Secretary encryption case — matthew_d_green · 2026-11-20
- PHP bug sent TLS-protected data in plaintext if you capitalized the S in https:// — matthew_d_green · 2026-11-20
- Joshua Saxe briefs Congress: fix vendor security liability, craft nuanced open-weights policy — joshua_saxe · 2026-11-20
- Superforecasters put 33% odds on a week-long outage hitting 1M+ people in an OECD country by 2028 — NathanpmYoung · 2026-11-20
- Sentinel brief 'Consider Cyberarmageddon' extrapolates cyberattack growth to societal failure — NathanpmYoung · 2026-11-20
- Protest at Anthropic HQ over banning RSI meets counter-protester with 'AI=GOOD' sign — nmasc_ · 2026-11-20