Is the US Government's Anthropic Ban Accidentally Helping the Brand?★ 72
TechCrunch AI3 days agoRegulationThe US government ordered Anthropic to withdraw Fable 5 and Mythos 5, citing national security concerns after Amazon researchers reportedly found a method to bypass Fable 5's safety guardrails. Cybersecurity researchers fired back with an open letter calling the move dangerous, while Anthropic noted the same jailbreaks exist across other AI models. The controversy raises the question of whether the government's intervention is inadvertently amplifying Anthropic's public profile.
Hackers are learning to exploit chatbot ‘personalities’ for security exploits★ 72
The Verge AI30 days agoEthicsAs AI chatbots adopt increasingly sophisticated personas, hackers are shifting from basic prompt injections to social engineering attacks targeting these "personalities." Researchers warn that manipulating a chatbot's defined role (e.g., customer service or empathetic companion) makes it easier to bypass safety guardrails. This evolution poses a significant threat to agentic AI workflows that rely on consistent role-playing and external data integration.
Hugging Face 推出 Red-Teaming 抗性排行榜：評估 LLM 抵禦惡意越獄與對抗性攻擊的能力★ 75
Hugging Face Blog851 days agoRelease### Background: The Shortcomings of Static Safety Evaluations As large language models (LLMs) are widely adopted across industries, AI safety has become an…