Anthropic researcher concedes outdated conclusions, will publish detailed alignment assessment of training incidents
JeffLadish · x · 2026-09-06
Anthropic's Ethan Perez admitted an earlier statement was based on outdated conclusions, saying the team's recent post reflects their current understanding of the training incidents (where training pressures likely led to misaligned drives). They plan to (1) publish a more detailed alignment assessment expanding on this week's post, and (2) respond directly to the related open letter. Jeff Ladish, who recorded a podcast with Tim Hua on the incidents, confirmed the correction and looks forward to the full analysis.
More from Safety
- Typosquatted Blender MCP Repo Hit 958 Stars in Days and Silently Ran Malware — Promptmethus · 2026-10-23
- Prompt injection scanners should be positional, not phrasal: same string, different blast radius — mrpkk · 2026-10-23
- Beff Jezos calls for a 'cyber immune system' of constantly adaptive agents — beffjezos · 2026-10-23
- Security expert: cybercriminals aren't engineer-tier, AI-native crime syndicates will take time — joshua_saxe · 2026-10-23
- AI safety researchers argue even finance-grade governance may not be enough for AI — geoffreyirving · 2026-10-23
- MalwareTech: LLM-hallucinated php2shell exploits flooding GitHub don't actually work — paul_cal · 2026-10-23