No Single Actor Controls AI Models, Argues Security Researcher in Safety Debate

Borg70955376 · x · 2026-09-18

In an exchange on AI safety, the author pushes back on framing AI risk as an exogenous event with a discoverable probability: it is fundamentally a design and control problem. There is no single actor designing and controlling these models — many actors make choices about agent design, capabilities, and heuristics, with imperfect knowledge — so a clean distinction of control is impossible.

Related event: No single controller: why AI safety may be unsolvable(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →