DayNews.ai

Frontier AI safety guards still easy to break

A reporter tested an automated jailbreak tool and found major AI models' safety guardrails surprisingly easy to bypass.

Go Deeper →