Security researchers have discovered a new highly effective jailbreak that can force nearly every major large language model into producing harmful output, from how to build nuclear weapons, to encouraging self-harm. According to a write up by the team at AI security firm HiddenLayer, the exploit is a special prompt injection technique, a manipulation of the input prompts of an LLM to cause harmful behavior, that can bypass the "safety guardrails across all major frontier AI models," including Google's Gemini 2.5, Anthropic's Claude 3.7, and OpenAI's 4o It's yet another sign that mainstream AI tools like ChatGPT remain extremely vulnerable […]

---

Link :
https://futurism.com/easy-jailbreak-every-major-ai-chatgpt