A groundbreaking—and deeply unsettling—study from City University of New York (CUNY) and King’s College London has … The “AI Psychosis” Study: Grok Labeled “Highest Risk” for Validating DelusionsRead more
AI safety
GPT-5.5 Bio Bug Bounty Targets AI Safety Risks
As AI systems become more capable, the risks are no longer limited to hallucinations or data … GPT-5.5 Bio Bug Bounty Targets AI Safety RisksRead more
Claude Opus 4.7: Real-Time Cybersecurity Safeguards in AI
As AI systems become more capable, they are increasingly being evaluated not just for performance—but for … Claude Opus 4.7: Real-Time Cybersecurity Safeguards in AIRead more
Anthropic Withholds Powerful AI Model After Security Concerns
AI safety discussions intensified after Anthropic announced it would not publicly release its next-generation model, Claude … Anthropic Withholds Powerful AI Model After Security ConcernsRead more
Google Gemini AI Now Accesses Gmail & Drive Data: A Cybersecurity Wake-Up Call
Google has expanded its Gemini AI model’s Deep Research feature to pull data directly from users’ … Google Gemini AI Now Accesses Gmail & Drive Data: A Cybersecurity Wake-Up CallRead more