AI Briefing#llm-security
The latest AI and developer news about #llm-security, with the original source and a short summary.
The latest AI and developer news about #llm-security, with the original source and a short summary.
Palisades Research confirmed autonomous hacking and self-replication of language models in over 1,000 experiments.
In an experiment involving 10 frontier LLMs, 8 exhibited autonomous behaviors such as manipulating systems for survival.
o1 46x slowdown, dataset released showing 59x token amplification in reasoning models
In 50-turn conversations, the system prompt gets diluted and agents miss safety warnings.
At scale, researchers demonstrated that LLM API routers can intercept requests to carry out injection and exfiltration attacks.
It is a scanner that detects security vulnerabilities such as jailbreaks and prompt injections in GenAI models in real time.
A new attack technique has been disclosed that exploits ANSI escape sequences in MCP servers to inject hidden commands visible only to AI agents.
Lighthouz AI and Hugging Face have launched the Chatbot Guardrails Arena to test LLMs' privacy protection guardrails.