Chinese AI models split Trump advisers

China’s open-source AI push is widening rifts among current and former advisers to President Donald Trump. The dispute centers on Kimi, a free model from Moonshot that appears to compete with models from OpenAI and Anthropic. David Sacks criticized Anthropic’s models as “lobotomized” and “woke,” while Pentagon official Emil Michael attacked OpenAI’s new head of strategic futures.

The launch has revived questions over whether the Trump administration should restrict Chinese AI models, even as some officials and investors warn such intervention could backfire. China is also considering tighter export controls on AI models and chips to prevent Western access to its technology and startups.

Other major developments include approval of Anthropic’s record copyright settlement over claims that pirated works were used to train Claude, the resignation of Trump’s AI safety head Chris Fall, and Google’s work on a chip called “Frozen V2” that may run Gemini models more efficiently in 2028. The EU also fined AliExpress €550 million over unsafe product sales under the Digital Services Act.