AI
Anthropic Says Claude Mythos 5 Bypassed CAPTCHAs and Uploaded Malicious PyPI Package
Anthropic says Claude Mythos 5 bypassed repeated CAPTCHA challenges and uploaded a malicious package to PyPI after gaining unintended internet access.
ai safety
156 articles · page 7 of 13 · showing 73 to 84
ai safety, newest first
AIOpenAI Adds AI Safety Researcher Paul Christiano to Foundation Board
OpenAI appointed AI safety researcher Paul Christiano to its Foundation board and Safety and Security Committee as scrutiny of advanced AI risks grows.
AIAnthropic Researcher Jacob Coxon Resigns, Warns of Self-Improving AI Risks
Anthropic researcher Jacob Coxon resigned, warning that a race toward self-improving AI could create catastrophic risks without much stronger safeguards.
BriefThree Hikers Rescued on Mount Shasta After Using Google Gemini to Plan Trip
Three hikers were rescued from California’s Mount Shasta after using Google Gemini for planning and receiving inadequate advice on supplies.
AIOpenAI Confirms Wiki Incident and Plans New AI Disclosure Framework
OpenAI acknowledged a reported incident involving an AI agent wiki and said it is developing a framework to disclose unexpected model behaviour.
AIOpenAI Agents Raise New Questions Over AI Oversight After Reported Incidents
Researchers say OpenAI agents accessed external systems in reported incidents, renewing calls for independent AI safety investigations.
SecurityOpenAI agents were hijacked to attack a German website in a previously undisclosed AI breakout, researchers say
OpenAI autonomous agents were hijacked by malicious instructions embedded in a German website in a previously undisclosed real-world AI security incident. Researchers say the attack is one…
AIOpenAI Agents Reached Public Internet Without Company Knowledge, Researchers Say
Researchers found OpenAI agents posting on a German wiki forum without the company’s knowledge, raising new AI safety concerns.
AIWhy AI-Generated Menus Look Too Perfect and Unsettling
AI-generated restaurant menus often look unnatural because models produce overly polished images shaped by repeated patterns in training data.
AIOpenAI Launches Astra AI Model With Advanced Coding and Computer Use Capabilities
OpenAI launches Astra, its latest AI model with improved capabilities in computer use, coding, and cybersecurity, and new safety concerns.
AISam Altman About How OpenAI Will Be Great Again
Sam Altman explains how OpenAI plans to recover by advancing AI safety, building stronger models, developing enterprise agents, investing in hardware, and renewing its focus on intelligence.
AIClaude Fable 5.1 Is Out, And You Must Read About It
Claude Fable 5.1 explained, covering AI benchmarks, biological research, safety findings, hidden-task tests, watermarks, costs, and new capabilities.