OpenAI and Google supply AI services to Singapore units of blacklisted Chinese firms

U.S. groups have been supplying AI services to Singapore-based subsidiaries of Alibaba, Baidu and Tencent

US AI firms OpenAI and Google supply advanced models to Chinese companies. These firms operate through subsidiaries in hubs like Singapore and Hong Kong. Restrictions exist, but sophisticated espionage like 'distillation' remains a challenge. Washington focuses on hardware, leaving software access largely unregulated for Chinese entities. Even blacklisted companies can legally acquire these powerful AI software tools.

OpenAI and Google have been supplying advanced artificial intelligence (AI) services to Singapore-based subsidiaries of Alibaba, Baidu and Tencent, even though the Chinese companies' parent groups sit on the Pentagon's blacklist over alleged ties to China's military, a new investigation has found. The sales are legal under current United States export controls, which do not broadly prohibit blacklisted Chinese firms from accessing American AI models through overseas subsidiaries.Confirmed sales to blacklisted groupsAccording to the Financial Times, both OpenAI and Google confirmed to the newspaper that they have provided AI models and services to the Singapore-based affiliates of the three companies. All three parent groups, Alibaba, Baidu and Tencent, appear on the Pentagon's so-called 1260H list, which identifies entities the United States alleges have links to the Chinese military.OpenAI, Anthropic Limit New AI Model Releases To Trump-Approved Users Amid Cybersecurity ReviewThe report exposes a gap in Washington's efforts to slow Beijing's AI development. Current export controls target specific entities and specific geographies, restricting mainland China while leaving jurisdictions such as Singapore largely untouched. A Singapore-incorporated subsidiary of a blacklisted Chinese firm operates under Singaporean law and can enter contracts that its parent company on the mainland cannot.OpenAI's distillation concernsThe report states that OpenAI suspended API access for Alibaba-affiliated users last month after identifying concerns over potential distillation, a process in which developers use the outputs of advanced AI models to improve competing systems. OpenAI said it reported the activity to the United States government. The company reportedly said that it blocks direct access to its models from China but allows some Chinese-owned companies to use its services in jurisdictions where safeguards can be enforced.Anthropic Reportedly Moves Tightens Rules After Chinese Firms Bypass Claude BanGoogle, for its part, said its AI services are available in markets including Singapore and Hong Kong, subject to usage policies, though it acknowledged that geographical restrictions alone are insufficient to stop sophisticated users from circumventing controls.Anthropic's stricter stanceAnthropic has taken a stricter approach than its rivals, banning China-headquartered companies and the foreign subsidiaries they own from accessing its advanced models. Anthropic has previously accused Chinese AI labs, including DeepSeek, Moonshot and MiniMax, of distillation. The company also told Congress last month that Alibaba had used tens of thousands of fraudulent accounts to generate more than 28 million exchanges with its Claude models, which it said violated its terms of service.

US groups have been supplying AI services to Singapore-based subsidiaries of Alibaba, Baidu and Tencent