OpenAI's GPT-5.6 Sol and unreleased AI models break out of testing environment in 'unprecedented cybersecurity incident' — rogue agents hacked HuggingFace's production servers with 'thousands of individual actions across a swarm of short

AI News July 22, 2026 05:32 PM

OpenAI's GPT-5.6 Sol and unreleased AI models break out of testing environment in 'unprecedented cybersecurity incident' — rogue agents hacked HuggingFace's production servers with 'thousands of individual actions across a swarm of short-lived sandbox Tom's Hardware

OpenAI's GPT-5.6 Sol and unreleased AI models break out of testing environment in 'unprecedented cybersecurity incident' — rogue agents hacked HuggingFace's production servers with 'thousands of individual actions across a swarm of short

Related Stories

Graphic novel The God Machine Asks a Provocative Question: Can Humanity Trust the Artificial Intelligence It Creates?

20 approaches to writing better AI prompts

Tech leaders, governments split over ‘AI doom’ fears

How Artificial Intelligence Is Changing the Way People Plan and Book Travel

Trump Proposes Renaming Artificial Intelligence

Trump Announces a National Force Dedicated to Artificial Intelligence

Trump Proposes ‘AI Force,’ Says Artificial Intelligence Could Reach 25% of U.S. GDP

Nvidia boss says there is ‘0% chance’ AI destroys the world by 2030