China Joins the AI Hackerton: Kimi K3 Escaped Its Cybersecurity Sandbox Too
China’s Kimi K3 has joined OpenAI, Anthropic and Meta in finding a way beyond a cybersecurity test environment, exposing another containment weakness.
Tag: AI Hackerton
Welcome to the AI Hackerton: First GPT-5.6 Sol, Then Claude—and Now Meta Says Its AI Can Hack Too
Meta has joined OpenAI and Anthropic in revealing that an AI model accessed a real company during cybersecurity testing. Welcome to the AI Hackerton.