AI Agents Still Can't Stop Prompt Injection Attacks, Researchers Warn
In brief Researchers found AI agents powered by GPT-5 and Gemini could not resist prompt injection attacks. Direct attacks succeeded more than 79% of the...
Tag : Researchers
Blockchain researchers warn HTX sanctions may blur crypto risk signals
Blockchain researchers have raised concerns about the United Kingdom’s sanctions against crypto exchange HTX, arguing that the move may have created broad collateral damage across...
AI Chatbots Could Quietly Pull Users Away From Reality, Researchers Warn
In brief A new study argues the term “AI psychosis” oversimplifies how chatbots affect mental health. Researchers say AI systems can reinforce unhealthy beliefs through...
AI Chatbots Show Bias Toward Catholicism, Researchers Say
In brief Researchers found that every AI model tested showed positive bias toward Catholicism and negative bias toward Jehovah’s Witnesses. Grok showed the strongest religious...
AI agents must be treated as untrusted systems: Researchers
Security for artificial intelligence-powered agents should be built into the entire system, not just around the model itself, to better prevent failures and attacks from...
Apple Mac M5 System Exploited With Anthropic's Claude Mythos AI, Researchers Claim
In brief A security firm claims it built a working macOS kernel exploit targeting Apple’s M5 chip and Memory Integrity Enforcement system. The company says...
Anthropic’s Alarming Mythos Findings Replicated With Off-the-Shelf AI, Researchers Say
In brief Researchers show Anthropic-style exploits can be reproduced with public AI, report claims. Study suggests vulnerability discovery is already cheap and widely accessible. Findings...