A bombshell revelation is putting the US government in an extremely difficult position, exposing a massive "loophole" in the tech war with China. As reported by the Financial Times, OpenAI and Google are selling artificial intelligence services to subsidiaries of Alibaba, Baidu, and Tencent in Singapore. The fact that these parent groups are included in the Pentagon's notorious 1260H list as Chinese military collaborators has triggered a storm of reactions in Washington, with experts warning that China is systematically "stealing" cutting-edge US expertise without spending a single dollar.
Reactions within the US
Although the sales are legal, they have reignited calls for stricter regulation of US AI models, similar to the restrictions in place for the export of microchips used to train powerful models. The US government has moved toward controlling access to individual advanced AI models, such as Anthropic's Mythos and Fable and OpenAI's GPT-5.6. However, it has not banned the general use of cutting-edge AI software by China-based entities, even those included in the so-called 1260H list, a "blacklist" mandated by Congress that includes Chinese companies with alleged links to the People's Liberation Army.
"The [Trump] administration keeps saying we need to beat China in artificial intelligence, but the problem is they haven't done anything regarding export controls, which is the actual tool we have to slow China down," said Chris McGuire, a technology and security expert at the Council on Foreign Relations.
The recent blocking of Alibaba users by OpenAI followed suspicions of "distillation", a process where developers use the outputs of AI models to improve competing systems. This activity was flagged to the US government, a spokesperson said. OpenAI stated that it does not allow access to its models within China, but confirmed that it permits "certain companies" of Chinese ownership or based in China to use its tools for operations in "countries where we can enforce safeguards and monitor for instances of distillation." The company added: "We prefer to see more of the world using artificial intelligence shaped by democratic values, rather than AI controlled by authoritarian governments." It further noted that "we do not believe that nationality alone should determine access."
"Frontier AI models should not be provided to China-based companies anywhere in the world," added McGuire, who worked on export controls as a senior White House and State Department official during the Biden administration. Anthropic has previously accused Chinese AI labs DeepSeek, Moonshot, and MiniMax of distillation. Last month, it warned in a letter to Congress that the Chinese e-commerce giant Alibaba used 25,000 fraudulent accounts to generate more than 28.8 million interactions with Claude, which it stated constituted a violation of its terms of service.
www.bankingnews.gr
Σχόλια αναγνωστών