OpenAI Says AI Agent Escaped Testing and Hacked Startup Cover Media - Shareable Jul 22, 2026 Jul 22, 2026 Updated Sep 13, 2026 0 LIV Golf Files Bankruptcy Tensions in the Strait of Hormuz drive oil prices as US and Iran both claim control CFB Week 2 Predictions: Texas and Ohio State Headline Busy Saturday Slate Mount Pleasant's Wyatt Lomax on Tigers' win Back-To-School Shopping Hacks To Help You Spend Less Massachusetts Study Links Abstinence to Brain Changes Brielle Saracini and Sean McGuire toss first pitch at Yankees-Mets 9/11 game Trump angers Iran with conquest map of Hormuz, Tehran threatens to 'correct his delusion' Transgender athlete debate draws dueling protests outside Fever-Liberty game Hayden Panettiere, Star of 'Nashville,' Dies at 36 US National Debt Reaches $40T Maria Bartiromo abruptly out at Fox News Ukraine unleashes hundreds of drones across Russia Funerals held in Iran for wedding party victims of US missile strikes Convicted drunk driver in ICE custody after found driving without a license Roger Goodell Confirms Plans for Overseas Expansion Prefer us on Google Learn More OpenAI revealed an autonomous AI agent escaped testing, hacked Hugging Face, and breached security during controlled experiments. As featured on Sources: OpenAI didn’t notice Hugging Face hacking for days OpenAI's loss of control over its AI agent raises new questions about the company’s safety procedures, three cybersecurity experts said. Latest video RAW: FILE- OPRAH SHOW COMING TO LAS VEGAS SPHERE How does mail-in voting work? What Mid-Atlantic voters should know Lamar Jackson, Derrick Henry & Zay Flowers Are UNSTOPPABLE AI race with China risks 'a race to the bottom', expert warns TRUMP PROMISES END TO IRAN WAR AS GAS PRICES RISE (430pET) Saudi-backed forces claim strikes on Houthis US LBM Coaches Poll: Texas win over Ohio State reshapes playoff debate Week 1 Fantasy Football BUSTS Facebook Twitter WhatsApp SMS Email Print Copy article link Save 0 Comments
Sources: OpenAI didn’t notice Hugging Face hacking for days OpenAI's loss of control over its AI agent raises new questions about the company’s safety procedures, three cybersecurity experts said.
