Moonshot limits the hacking potential of its Kimi model.

Chinese AI company Moonshot has limited the offensive hacking capabilities of its Kimi K3.1 Pro model by removing cybersecurity-related data from its training set, aiming to reduce misuse risks while retaining general programming functions. The company highlights the challenges of managing open-weight AI models, which increase both opportunities for security audits and risks from uncontrolled use, especially in securing critical industrial infrastructure.

This Wire brief sits within Fusion42's coverage of Cybersecurity, and 2 sources have reported it.