François Chollet: don't hand critical processes to autonomous AI agents, ever
fchollet · x · 2026-09-04
Keras creator François Chollet argues that even if AI develops the capability for advanced autonomy, we should not make it highly autonomous.
- We need a deliberate effort to keep humans in the loop across all critical economic and social processes, regardless of technical necessity
- Maintain control, visibility and understanding of critical processes; don't blindly hand everything to AI agents just because we can
- "AI as a tool in the human hand is the only form of AI worth pursuing"
Quoted @ccatalini adds: to correctly represent human preferences, humans must stay in the loop — via augmentation or better tooling.
Related event: Chollet: keep humans in the loop even if AI doesn't need them(4 posts)→
More from AGI Musings
- Why self-modifying AI is the sharpest test case for rational choice theory — jessi_cata · 2026-09-04
- Stewart Alsop: Most People Get LLMs Wrong, From Hype to Dismissal — StewartalsopIII · 2026-09-04
- Rational choice theory only becomes true in the superintelligence limit, argues poster — jessi_cata · 2026-09-04
- davidad on AI alignment and well-being: a truly aligned mind's welfare lies in the Good — mimi10v3 · 2026-09-04
- NYT Bestselling AI Book Got Data Center Water Usage Wrong by 4,500x — thursdai_pod · 2026-09-04
- Yacine: most people can't tell a Google search from an LLM, or grasp the TikTok algorithm — yacineMTB · 2026-09-04