normative

Closed-loop recursive self-improvement should be illegal

The single most important guardrail is banning closed-loop recursive self-improvement—where an AI runs its own experiments and rewrites itself a million times—because that is an event horizon with no way to know what comes out the other side; this should be illegal with jail time, since there is no reason to believe anyone could do it safely.

normativepending

Speaker

Tristan Harris

Evidence Quote

“that should be illegal and there should be jail time if you do that”

Source

469 - Escaping An Anti-human Future— Making Sense with Sam Harris
Created: 6/14/2026, 2:01:37 AM

Your notes

Loading notes...