AI: Moonshot curbs hacking abilities of latest model
Moonshot AI is restricting its own AI model. Kimi K3.1 Pro is supposed to be less able to hack. To achieve this, cybersecurity data was removed from the training.
Moonshot AI, a Chinese start-up based in Peking, has limited the hacking capabilities of its latest AI model, Kimi K3.1 Pro. The company removed cybersecurity-related data from its training datasets to restrict the model's attack abilities without affecting its general programming capabilities. Moonshot AI describes Kimi K3.1 Pro as its most powerful open-weight model, trained with three trillion parameters.
The company plans to release the model's weights by November 4, allowing external developers to execute and modify the model on their own systems. This development comes as the Chinese government proposes export controls on AI-powered technology for generating and modifying malware.
Written by urgent.news from Handelsblatt's report — not a translation of it. Machine-written — may contain errors; check the original before relying on it.
