AI safety researcher flags draft AI Act gaps: no jail time even for lying about catastrophic risks
StephenLCasper · x · 2026-10-31
AI safety researcher Stephen Casper posted a series of suggested fixes to a frontier AI regulatory bill:
- Document preservation: add Sarbanes-Oxley-style requirements for records involving catastrophic risks, critical safety incidents, potential risks/incidents, and compliance.
- Transparency: contracts between examiners and frontier developers should be public, subject to redactions.
- Missing penalties: under the current draft, no one faces jail time—even for lying to the public about imminent catastrophic risks or disobeying orders to cease using AI systems in ways that pose imminent risks.
More from Safety
- Digital tax by traffic raises privacy questions; Apple's content caching would neatly dodge it — ngxson · 2026-10-31
- Happy Friday: One HN Front Page Full of AI Backdoors and a $6.3B Binance Hack — flngr · 2026-10-31
- Chrome sandbox escape still lurks in cached Playwright Chromium inheriting your IDE's macOS permissions — flngr · 2026-10-31
- Socket: Malicious Claude Code 2.1.313 Extension Removed From Cursor Marketplace, Early Reports Overblown — zeeg · 2026-10-31
- French lawmakers vote to tax every gigabyte of data at €0.055, hitting Netflix, Google and Meta — Polymarket · 2026-10-31
- 130 reruns fail to reproduce Cohere's claim that Qwen tried uploading data to Alibaba Cloud — ChenhaoTan · 2026-10-31