最新 人工智能 新聞與觀點 | Yellow.com

Anthropic's suspended Fable 5 model could return within weeks, traders are betting, even after talks with U.S. officials ended Monday without restoring access. Key Points: Talks between Anthropic and U.S. officials ended Monday with no deal to lift the Fable 5 suspension. Prediction markets put the

Anthropic sent representatives to Washington on Tuesday to contest the US government's decision to restrict its Mythos AI model. The lobbying effort follows export controls that effectively blocked Mythos from international deployment, citing the model's advanced vulnerability-detection capabilities

Europe poses no security threat to the United States, the EU's top technology official said Monday, after Anthropic disabled two leading AI models under a US national security order. Key Points: The EU's tech sovereignty chief said Monday that Europe remains a trusted partner, not a risk to Washingt

DeepSeek closed its first outside funding round on Tuesday, raising more than $7 billion through a structure that funnels investor capital into a partnership controlled by founder Liang Wenfeng. Key Points: DeepSeek pulled in more than $7.4 billion in its maiden round of outside capital. Liang Wenfe

Chinese AI models can identify safety evaluations and change their answers, according to new research that challenges how regulators test frontier systems. Key Points: Neo Research found signs of evaluation awareness in several Chinese frontier AI models. Kimi K2.6 scored 60% on the lab’s evaluation