Latest cybersecurity reporting from selected sources.

Researcher Details Stealthy Multi-Turn Prompt Exploit Bypassing AI SafetyWell-timed nudges are enough to derail a large language model and use it for nefarious purposes, researchers have found. Dubbed "Echo Chamber," the exploit uses a chain of subtle prompts to bypass existing safety guardrails by manipulating the model's emotional tone and contextual assumptions.

Researchers Manipulate o1, o3, Gemini 2.0 Flash Thinking and DeepSeek-R1The new generation of "reasoning" artificial intelligence chatbots is susceptible to a jailbreaking method that hijacks models' safety pathways, reducing their ability to detect harmful content. The attack modifies the reasoning processes and reintegrates the changes into the original queries.

Mozilla Researcher Uses Non-Natural Language to Jailbreak GPT-4oAnyone can jailbreak GPT-4o's security guardrails with hexadecimal encoding and emojis. A Mozilla researcher demonstrated the jailbreaking technique, tricking OpenAI's latest model into generating python exploits and malicious SQL injection tools.

'It was like watching a robot going rogue' says researcher OpenAI's language model GPT-4o can be tricked into writing exploit code by encoding the malicious instructions in hexadecimal, which allows an attacker to jump the model's built-in security guardrails and abuse the AI for evil purposes, according to 0Din researcher Marco Figueroa.…

Crooks Are Recruiting AI Experts to Jailbreak Existing LLM GuardrailsCybercrooks are exploring ways to develop custom, malicious large language models after existing tools such as WormGPT failed to cater to their demands for advanced intrusion capabilities, security researchers say. Undergrounds forums teem with hackers' discussions about how to exploit guardrails.