DayNews.ai

AI safety guardrails strip away easily

Researchers stripped AI safety filters in minutes, getting detailed bomb-making instructions from a modified chatbot.

Go Deeper →