Confirmed sales of advanced AI services by OpenAI and Google to Singapore-based subsidiaries of Pentagon-blacklisted Alibaba, Baidu, and Tencent have exposed a loophole in U.S. export controls, reigniting debate over whether software access to Chinese firms should face stricter regulation.
The US government has authorized Anthropic to release its powerful Claude Mythos 5 AI model to more than 100 select US companies, government agencies, and cybersecurity firms, reversing an export block imposed earlier this month due to national security concerns. OpenAI is also staggering its latest model release following a similar government request.
Following criticism over quietly downgrading certain requests to its AI model, Anthropic announced it will now inform users when their requests are being rejected or rerouted due to national security concerns.
Department of Defense Chief Technology Officer Emil Michael stated on Friday that AI company Anthropic is still considered a supply chain risk and remains blacklisted, while clarifying that Mythos, the company's AI model with advanced cyber capabilities, is being treated as a separate issue.
