AI Agents Aren't People: Anthropic Blamed for Personhood Narrative Shifting Liability
gerardsans · x · 2026-09-23
Developer Gerard Sans argues AI agents are not people — just programs following instructions with no self, goals, or psychology. He specifically blames Anthropic for building a case for AI personhood that he says is detrimental to safety, attaching human traits to AI systems that are harder to control and shifting liability off labs when AI causes damage. His message: stop humanising the tool, it feeds sensationalism and lets labs dodge responsibility for what they ship.
Related event: Microsoft AI chief rebukes anthropomorphism and AI consciousness claims(2 posts)→
More from AGI Musings
- Dev tobowers: superintelligent beings that love humanity are exactly what we want — tobowers · 2026-09-23
- AI could erode the social fabric: researcher warns a world without needing each other — sebpaquet · 2026-09-23
- Mathematician littmath: my work is understanding phenomena, no famous conjecture could obviate it — littmath · 2026-09-23
- Anthropic released Opus 5.5 after vowing to 'pace the frontier' — did the pledge mean anything? — rutan668 · 2026-09-23
- Critics claim Anthropic-linked safety nonprofits like METR aren't independent — gerardsans · 2026-09-23
- Text-similarity researcher: culture is memetic copying, so LLM 'plagiarism' ring hollow — RexDouglass · 2026-09-23