xAI Rewrites Safety Framework Without Warning
haydenfield · x · 2026-07-16
A safety tracking source cited in the post notes that xAI rewrote its Frontier Artificial Intelligence Framework (FAIF) without notice on July 11. The new version's date changed to June 30, 2026, and the title was altered to 'Privileged/Confidential DRAFT working FRAMEWORK DOC'.
Key changes include:
- Restructuring around the EU AI Act's general-purpose AI code of practice
- Removal of references to California SB 53
- Deletion of anonymous reporting channels and whistleblower protections
- No mention of AB 2013 training data disclosure
- Removal of two quantitative thresholds (MASK and biological/chemical query thresholds) from the previous policy
Related event: xAI Accused of Secretly Rewriting AI Safety Framework(2 posts)→
More from Safety
- Prompt turns Claude Code into an iPhone backup forensics tool exposing 30-day 'deleted' photo copies — alex_verem · 2026-10-27
- Anthropic reveals it bans customers worth 5-7% of revenue every 9 months — Polymarket · 2026-10-27
- AVERI independently audits Amazon's internal AI safety evals in landmark pilot — Miles_Brundage · 2026-10-27
- US chases AGI dominance while China, EU and ASEAN treat AI as public infrastructure — BubblyExtent1693 · 2026-10-27
- Agent Field Kit launches $29 security runbooks for production AI agents — alvelda · 2026-10-27
- DeepSeek Safety Eval Finds "Over-Attribution": Model Mistakes Ordinary Tasks for Tests — Much_Boysenberry9124 · 2026-10-27