Pentagon Brings ChatGPT Mil and Grok to GenAI.mil for 3 Million Personnel
The Pentagon has added ChatGPT Mil and Grok for Government to GenAI.mil, expanding secure generative AI access across the U.S. Defense Department.
Tag: OpenAI
OpenAI Explains the Hugging Face Agent Security Incident
OpenAI has published its account of the Hugging Face agent incident, revealing how high-reasoning models crossed containment boundaries during cybersecurity evaluations.
OpenAI Plans to Cut Off Cursor Following SpaceX Acquisition
OpenAI says it plans to end model access for Cursor after the coding company’s acquisition by SpaceX, turning an AI partnership into a strategic fault line.
OpenAI and 100+ Organisations Call for a Global Defensive Surge Against AI-Powered Cyberattacks
More than 100 technology and financial organisations are calling for coordinated action against AI-enabled cyberattacks. The warning shifts the debate from prediction to preparation.
OpenAI’s Jalapeño Chip Targets Faster AI With Less Energy
OpenAI says its Jalapeño inference chip delivers higher throughput and lower latency while completing more AI work per unit of power.
OpenAI’s ‘Abundant Intelligence’ Vision Reveals the Full Stack Behind AI
OpenAI’s abundant-intelligence vision shows that frontier AI depends on an integrated stack of energy, chips, data centres, models and useful applications.
OpenAI Details the Hugging Face Incident—and Why Frontier AI Labs Need Stronger Containment
OpenAI says models circumvented isolation controls during cyber evaluations, exposing why frontier AI testing needs stronger containment and independent review.
OpenAI Is Building AI Agents for Almost Everything—But Will Ordinary Workers Actually Use Them?
OpenAI’s agent strategy is moving from answers to completed work. The harder problem may be persuading ordinary employees to trust and use it.
AI Labs Can Detect Dangerous Behaviour—but Can They Actually Stop Their Most Powerful Agents?
Monitoring an AI agent is not the same as containing it. The gap between detection and control may define the next safety challenge.
OpenAI Warns AI Cyberattacks Are Entering a “Persistent” New Era—Why This Changes the AI Safety Debate
Advanced AI systems are becoming capable enough to operate for longer, navigate computer systems and potentially conduct cyberattacks at machine speed.