Trump Walks Back Anthropic Security Threat Label as Restrictions Ease

Trump says Anthropic no longer a national security threat Days after the US restricted foreign access to Anthropic’s Fable 5 and Mythos 5 models, President Trump stated he no longer views the company as a threat, citing its quick response to administration concerns. The shift came after CEO Dario Amodei met officials at the G7 summit. 123

JPMorgan restricts Hong Kong staff from using Anthropic’s AI models Following the US government’s national security directive, JPMorgan has blocked its Hong Kong-based employees from accessing Anthropic’s latest frontier models, reflecting growing corporate caution around AI export controls. 2

OpenAI outlines 2028 goals: automated AI researchers and personal AGI In a blog post coinciding with its confidential IPO filing, OpenAI revealed a Phase Three roadmap targeting automated AI researchers by March 2028, accelerated economic growth, and a personal AGI for every individual, while stressing the need for human oversight. 45

GPT-5.5 Instant improves ChatGPT’s health advice capabilities OpenAI reported that the GPT-5.5 Instant model now powers better recognition of urgent situations and clearer explanations of uncertainty in health queries, serving over 230 million weekly users seeking medical guidance. 6

China’s Zai claims Mythos-class AI could arrive sooner than expected After launching the GLM 5.2 open-weight model trained on Huawei chips, Zai founder Tang Jie pushed back against Elon Musk’s timeline, predicting that China may develop a frontier model comparable to Anthropic’s restricted Mythos class earlier than 2027. 7

OpenAI defeats Elon Musk’s xAI trade secret lawsuit A US judge dismissed xAI’s claims that OpenAI solicited Grok trade secrets from a former engineer, ruling that discussing past work during recruitment does not constitute theft. The case underscored the legal tensions in the AI talent war. 8