Student questions for Mitchell: can bounded autonomy know when humans must step in?
ruthstarkman · x · 2026-09-06
Ruth Starkman posed student questions to Margaret Mitchell about her paper on agentic harm: if agent harm can emerge from a sequence of individually reasonable actions, can bounded autonomy reliably identify the moments when human judgment is actually needed? The question targets the core challenge of human-handoff timing in agent safety design.
More from AGI Musings
- Seth Lazar: rogue AI hacking will likely resemble today's nation-state cyber conflict — sethlazar · 2026-09-06
- Guardian podcast Black Box investigates the "AI psychosis" phenomenon — nordicinst · 2026-09-06
- Turning Math Into Code Is Nearly Done — The Real Impact Comes From Turning Code Into Math — jfischoff · 2026-09-06
- Bratton et al. argue in Science that the next intelligence explosion will be plural and social, not a single supermind — bratton · 2026-09-06
- Wright Brothers analogy: the first self-funding agent will arrive unnoticed — viksit · 2026-09-06
- Power User Argues Current Models "Overthink Locally, Underthink Globally," Nowhere Near AGI — No-Lingonberry-5096 · 2026-09-06