AI governance scholar Luiza Jarovsky calls for a global ban on RSI

LuizaJarovsky · x · 2026-11-19

AI governance researcher Luiza Jarovsky states bluntly that recursive self-improvement (RSI) "will probably be very very harmful for humanity" and that the world should agree to ban it.

RSI — AI systems improving themselves and accelerating further capability gains — is a core AI safety concern. The statement reflects a hardline stance from an influential voice in AI policy discourse.

Original post →

More from AGI Musings

AGI Musings channel →