America is locked in a global race for artificial intelligence dominance, yet some of the country's most powerful AI companies have been supplying advanced technology to subsidiaries of Chinese firms that the Pentagon has identified as having ties to Beijing's military.
The disclosures, first reported by the Financial Times, are fueling fresh calls for the Trump administration to close what many national security experts view as a dangerous loophole in U.S. export policy.
According to the report, Google and OpenAI have provided AI services to Singapore-based affiliates of Alibaba, Baidu, and Tencent. All three Chinese technology giants appear on the Defense Department's Section 1260H blacklist because of their alleged connections to China's military.
The transactions are legal under current U.S. regulations. But legality is not the same as sound national security policy—especially when America's chief geopolitical rival is aggressively pursuing AI superiority.
OpenAI acknowledged that it recently suspended API access for Alibaba-affiliated users after detecting suspected "distillation," a technique in which developers use the output of one AI model to improve competing systems. The company said it also reported the activity to the U.S. government.
OpenAI emphasized that its models are blocked inside mainland China but said it allows "some companies" with Chinese ownership or headquarters to use its AI tools in countries where it believes safeguards can be enforced.
"We would rather see more of the world using AI shaped by democratic values than AI controlled by autocratic governments," the company said, adding that it does not believe "nationality alone should decide access."
Google defended its own policies, noting that its AI services are available in jurisdictions such as Hong Kong and Singapore while prohibiting distillation and other forms of misuse. The company argued that geographic restrictions alone are insufficient because sophisticated users can often work around them.
Those explanations have done little to calm concerns among security experts who argue that America's AI leadership cannot be protected if cutting-edge technology remains accessible through overseas affiliates.
The debate comes as the Trump administration continues emphasizing that the United States must maintain its technological advantage over Communist China. While Washington has already imposed sweeping restrictions on advanced semiconductor exports, comparable controls governing access to frontier AI models remain far more limited.
Current rules regulate access to certain advanced systems, including Anthropic's Mythos and Fable models and OpenAI's GPT-5.6. However, they do not broadly prohibit Chinese-headquartered companies—including those on the Pentagon's blacklist—from accessing advanced AI services outside mainland China.
Chris McGuire, a technology and security expert at the Council on Foreign Relations, argued that the gap leaves one of America's most important strategic tools sitting on the sidelines.
"The 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," McGuire told the Financial Times.
Anthropic has taken a far more restrictive approach.
The AI company prohibits Chinese firms and foreign entities they own from accessing its advanced models and recently tightened those restrictions after identifying what it described as attempts to circumvent its safeguards.
Anthropic has previously accused Chinese AI developers—including DeepSeek, Moonshot, and MiniMax—of engaging in distillation. In a letter to Congress last month, the company alleged that Alibaba created approximately 25,000 fraudulent accounts to generate more than 28.8 million interactions with its Claude AI model in violation of its terms of service. Alibaba did not respond to requests for comment on those allegations.
Meanwhile, Alibaba is challenging its designation on the Pentagon's blacklist, asking a federal court to remove what it called an "arbitrary and capricious" classification. Baidu declined to comment on the report, while Tencent did not respond to requests regarding its use of U.S. AI models.
National security specialists argue that the stakes extend well beyond corporate competition. Joe Khawam, an AI policy and national security law expert at the Law Reform Institute, has warned that stronger export controls are needed to prevent Chinese firms from acquiring advanced AI capabilities without paying the enormous research and development costs that American companies have invested.
The broader question is no longer whether AI will define the next generation of economic and military power—it almost certainly will. The question is whether the United States will allow companies tied to America's chief strategic adversary to benefit from American innovation while Washington races to stay ahead. For an administration committed to an America First agenda, that is a national security debate that is becoming increasingly difficult to ignore.
