Ex-DeepMind, now Anthropic researcher: no viable scientific plan for self-improving AI risk
asusarla · x · 2026-09-11
Anna Wang, formerly of Google DeepMind and now at Anthropic, stated in a personal capacity that there is not yet a viable scientific plan to solve risks from recursively self-improving AI — a sentiment she says is common among her peers.
Hugging Face's mmitchellai amplified the warning, calling on agent developers to make human oversight design central to AI agent development. The exchange highlights growing internal-lab concern about recursive self-improvement risk.
Related event: Researcher: No Scientific Plan Yet for Recursive Self-Improvement Risk(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