Several prominent U.S. data companies that support artificial intelligence development for American entities, including the Pentagon, are also providing services to Chinese AI labs, according to a report. These companies collate, label, and package vast amounts of information essential for training AI models.
Representative Michael McCaul (R-Texas), a former chairman of the House Foreign Affairs Committee, has raised concerns about this dual engagement. "America is in an AI race with the Chinese Communist Party," McCaul stated. "We cannot spend billions trying to protect our technological advantage while American companies package up our best human expertise and sell it to companies tied to the Chinese military."
China's top AI labs reportedly spend approximately $500 million annually on data services from American companies. Four U.S.-based firms identified as having done business with Chinese AI labs are SurgeAI, Mercor, AfterQuery, and Turing.
McCaul has indicated that companies seeking U.S. government contracts should cease working with Chinese entities designated as military companies by the Pentagon. "If you want to do business with the US government, you should not also be helping a Pentagon-designated Chinese military company build better AI," he said.
AI models require extensive, specialized, and accurately labeled data to refine their pattern prediction capabilities. The process of compiling and cleaning this data is notably complex.
SurgeAI, a leading data labeling company, has reportedly worked with Tencent, a Chinese firm involved in AI and video games. Although SurgeAI has secured small contracts with the U.S. Army and Air Force, Tencent is designated by the U.S. as a Chinese military company. Executives at Tencent have privately acknowledged their reliance on SurgeAI for AI development.
Mercor, which recently secured a U.S. government contract, has also reportedly engaged in business with Tencent. A source indicated that approximately 2% of Mercor's revenue in the second quarter came from China. The company specializes in recruiting and organizing experts for AI firms to train their chatbots.
AfterQuery, a company that curates data for AI development, is reported to generate $50 million in recurring revenue from China, including dealings with Ant Group, the financial arm of Alibaba. Turing, which assists in labeling and developing specialized training data, has worked with ByteDance, the parent company of TikTok in China.
The United States and China are engaged in a significant competition in the field of artificial intelligence. The U.S. has previously attempted to restrict China's access to advanced semiconductors crucial for AI development. However, access to meticulously processed American AI training data could significantly aid China's progress in the AI race. 