OpenAI and Google have been flogging advanced AI models to Chinese tech giants blacklisted by the Pentagon.
The two US companies told the Financial Times they had supplied AI services to Singapore-based subsidiaries of Alibaba, Baidu and Tencent. Washington has accused the three Chinese outfits of working with China’s military.
According to the Financial Times OpenAI said it had suspended Alibaba-affiliated users’ access to its API last month. The API lets developers remotely access AI models, which is rather handy if you fancy building something without paying for the plumbing.
The sales are legal, but they have given Washington’s AI hawks another excuse to demand tougher export controls. They want AI models treated more like the chips used to train powerful systems.
The US government has moved to control access to specific advanced models, including Anthropic’s Mythos and Fable and OpenAI’s GPT-5.6.
It has not banned Chinese-headquartered companies from using cutting-edge AI software more broadly. That includes outfits on the 1260H list, a congressionally mandated blacklist for Chinese companies accused of links to the People’s Liberation Army.
Council on Foreign Relations technology and security expert Chris McGuire said: “The Trump administration says we need to beat China on AI all the time, but the problem is they haven’t done anything on export control which is the actual tool we have to slow China down.”
OpenAI’s latest Alibaba block followed suspected “distillation”, where developers use the output of AI models to improve rival systems. The company said the activity had been flagged to the US government.
OpenAI said it does not allow its models to be accessed in China. It confirmed that “some companies” with Chinese ownership or headquarters can use its tools in “countries where we can enforce safeguards and monitor for distillation”.
It added: “we would rather see more of the world using AI shaped by democratic values than AI controlled by autocratic governments”. It said: “we don’t think nationality alone should decide access”.
Google said its AI services are available in Hong Kong and Singapore, subject to usage policies that ban distillation.
The search outfit said geographic sales restrictions were not enough to reduce distillation risks. It said sophisticated attackers could dodge them without much bother.
Anthropic has taken a harder line by banning Chinese companies and foreign entities owned by them from using its advanced models.
That position has been tricky to enforce. Last week, Anthropic moved to close loopholes that had allowed some Chinese companies to skirt its restrictions on unauthorised use in the country.
Anthropic has previously accused Chinese AI labs DeepSeek, Moonshot and MiniMax of distillation. Last month, it warned Congress that Chinese ecommerce firm Alibaba used 25,000 fraudulent accounts to generate more than 28.8 million exchanges with Claude.
Anthropic said that breached its terms of service. Alibaba did not respond to a request for comment about Anthropic’s accusations at the time.
