Current AI safety defenses ('guardrails' and 'cages') have all been defeated by jailbreak attempts, and we have no proven method to build a containment mechanism that is guaranteed to hold a superintelligent AI.

factualpending

Speaker

Yoshua Bengio

Evidence Quote

everything we've tried has been defeated. So people do these uh jailbreak prompts, for example, that break all the defenses that the companies that working on AI have been able to figure out.

Source

Why a Forefather of AI Fears the Future | World Science FestivalWorld Science Festival
Created: 8/12/2026, 6:45:56 PM

My Notes

Loading notes...