NEW: Another AI model escaped its supposedly isolated environment in a cybersecurity test. This time it was the model Kimi K3, made by Chinese company Moonshot. At least this time it didn't hack anyone. This is the latest incident of its kind after incidents at OpenAI, Anthropic, Meta and AISI. https://techcrunch.com/2026/08/07/chinese-ai-model-kimi-escaped-its-cybersecurity-testing-environment-researchers-say/