In the News

In a recent Politico newsletter , TIP Policy Director Geoffrey Cain warns that while the White House is moving to crack down on Chinese AI distillation attacks, U.S. tech companies may be undermining those efforts. Cain argues that policymakers “cannot ignore” how firms like Microsoft could be enabling access to advanced AI systems through global infrastructure, pointing to reported loopholes in cloud services. The warning follows new disclosures that foreign actors have used proxy accounts to extract capabilities from U.S. AI models, a growing tactic in the U.S.-China tech race. ### Full statement: “The White House is right to sound the alarm on China’s industrial-scale theft of American AI. But if Washington is serious about stopping distillation, it cannot ignore the role that U.S. tech giants themselves are playing in handing China the keys. “Microsoft runs an AI research lab inside China that has trained a generation of the country’s top AI talent — including multiple engineers who went on to help build DeepSeek. Microsoft provides Chinese companies, including state-owned enterprises, access to OpenAI’s advanced models through its Azure joint venture, even as OpenAI claims to have banned China from using its products. And Amazon, Oracle, Microsoft, and other U.S. cloud providers have been selling remote access to restricted U.S. chips and frontier AI models to Chinese state-linked entities—a loophole in current export controls. “You cannot credibly fight Chinese AI theft while your own business model depends on enabling it. As part of this welcome effort, the Administration should scrutinize U.S. companies’ China dealings that may be facilitating the distillation it aims to prevent. American AI leadership does not mean handing our crown jewels to Beijing in the pursuit of short-term profits.”

The White House Office of Science and Technology Policy (OSTP) on Wednesday accused foreign entities — principally based in China — of waging “industrial-scale” campaigns to distill American frontier AI models, using tens of thousands of proxy accounts and jailbreaking techniques to extract proprietary capabilities. In a memorandum to executive departments and agencies (NSTM-4), OSTP Director Michael Kratsios announced that the Trump Administration will share threat intelligence with U.S. AI companies, help industry coordinate defenses, and explore measures to hold foreign actors accountable. The memo follows reports from OpenAI and Anthropic documenting distillation attacks by Chinese labs including DeepSeek, Moonshot, and MiniMax, and a House Select Committee on China investigation finding that Chinese AI labs are using fraudulent accounts and unauthorized API access to train on U.S. models. Tech Integrity Project Policy Director Geoffrey Cain offered the following statement: “The White House is right to sound the alarm on China’s industrial-scale theft of American AI. But if Washington is serious about stopping distillation, it cannot ignore the role that U.S. tech giants themselves are playing in handing China the keys. “Microsoft runs an AI research lab inside China that has trained a generation of the country’s top AI talent — including multiple engineers who went on to help build DeepSeek. Microsoft provides Chinese companies, including state-owned enterprises, access to OpenAI’s advanced models through its Azure joint venture, even as OpenAI claims to have banned China from using its products. And Amazon, Oracle, Microsoft, and other U.S. cloud providers have been selling remote access to restricted U.S. chips and frontier AI models to Chinese state-linked entities—a loophole in current export controls. “You cannot credibly fight Chinese AI theft while your own business model depends on enabling it. As part of this welcome effort, the Administration should scrutinize U.S. companies’ China dealings that may be facilitating the distillation it aims to prevent. American AI leadership does not mean handing our crown jewels to Beijing in the pursuit of short-term profits.” 

In a new piece for The Spectator , TIP Policy Director Geoffrey Cain argues that Western tech companies, universities, and governments are systematically enabling China’s military and surveillance capabilities, often within the bounds of the law. Citing a federal indictment involving diverted AI servers, Cain says the real issue is not just illegal smuggling but the much broader, legal flow of technology and expertise to China. From tailored chip sales by NVIDIA to research partnerships with Chinese military-linked institutions, he warns that “the front door has been open for years.” Cain also points to cases like Microsoft’s past use of China-based engineers on sensitive U.S. systems and talent pipelines from its Beijing lab into China’s AI sector as examples of systemic risk. The bottom line: enforcement actions may catch bad actors, but Cain argues the larger threat is structural, with commercial incentives continuing to outpace national security safeguards.

In a recent piece for The Washington Examiner , TIP Policy Director Geoffrey Cain argues that conflicts of interest and national security risks are being overlooked in Microsoft’s government cloud business. Following reports that senior Justice Department officials involved in approving Microsoft’s GCC cloud service later joined the company, Cain warns that clear security red flags are being ignored. “Before the DOJ authorized GCC High, there were plenty of public reasons to treat Microsoft’s China operations as a serious security concern,” he says, pointing to the company’s Beijing research lab, ties to Chinese surveillance firms, and AI research collaborations with a Chinese military university. He adds that “any one of those facts should have triggered harder questions about putting sensitive law enforcement data on Microsoft’s cloud,” and argues that together they should have led to far more rigorous scrutiny. Cain also argues that the Justice Department’s dual role as both a customer and regulator creates a “structural conflict of interest,” weakening oversight of cybersecurity risks tied to sensitive government systems. The bottom line: weak scrutiny and revolving door dynamics help entrench a system with known vulnerabilities, raising broader concerns about how the government evaluates Big Tech partners.

In a recent podcast appearance , TIP Policy Director Geoffrey Cain argues that major U.S. tech companies remain deeply dependent on China, despite mounting national security risks. Speaking on the American Compass Podcast , Cain explains that Silicon Valley’s long-standing bet on engagement with China has backfired, strengthening the Chinese Communist Party and helping build its surveillance and AI capabilities. He warns that companies continue to expose U.S. supply chains to coercion and technology transfer, pointing to efforts by firms like NVIDIA to maintain access to Chinese markets even as restrictions tighten. Cain argues the underlying problem is structural. Tech companies are incentivized to stay in China, even when doing so undermines U.S. security. 

Pro Publica recently published a report detailing how Microsoft, for nearly a decade, has used Chinese engineers to help maintain the US Department of Defense’s computer systems. According to the report, the Chinese engineers are overseen by officials who lack the technical knowledge to detect potential cyber threats. Notably, under Chinese law, all Chinese citizens are required to assist the Chinese government in intelligence and espionage operations if compelled. Tech Integrity Project Policy Director Geoffrey Cain offered the following comment: "For years, I’ve warned that Microsoft’s extensive business operations in China are threatening US national security. Today’s news may be the most egregious revelation yet. In recent years, China has launched widespread and successful cyber attacks against a slew of US government agencies by exploiting their Microsoft systems. Today’s news may explain why Chinese hackers have such a knack for cracking Microsoft products. It is past time that the Department of Defense and other US agencies exercise their procurement power to force Microsoft to make a change. There is simply too much at stake to enable US tech companies to continue putting profits over American security." You can read the full report , “A Little-Known Microsoft Program Could Expose the Defense Department to Chinese Hackers,” in Pro Publica, and you can see Geoff’s post on X here .

Imagine running your company out of a high-tech campus with armed guards at the gates, listening devices in the walls and a landlord who insists you turn over your blueprints to keep the lights on. Most executives would scramble for the exits. But for years, American tech companies made peace with that setup in China. Now, exposed by tariffs, they’re feeling the consequences of their devilish bargain. By inviting scrutiny, the tariffs are forcing Big Tech to confront the vulnerabilities they long ignored: their excessive reliance on Chinese manufacturing, their reckless partnerships with companies connected to an authoritarian government and their transfer of sensitive technology into the hands of America’s largest adversary.  The results of the stress test of tariffs? They revealed that when forced to choose between shareholder returns and geopolitical alignment, many companies opted for market access — until that access came with a higher price.

According to the New York Post, a t least four current DeepSeek employees previously worked at Microsoft Research Asia, three of whom spent at least five years as research interns at Microsoft. “If you’re working for Microsoft for five years at any other job, that’s enough time to climb into a mid-senior or senior position,” said Geoffrey Cain, policy director at the Tech Integrity Project, in a New York Post article about Microsoft’s AI research lab in China . “Your typical intern does not spend five years developing AI software with that title, only to go on and to help release one of the most successful AI releases of all time.” “This is a deeper problem,” Cain added. “It’s that Microsoft is handing the training and the technology to China. It’s not necessary for these software developers in Microsoft to literally hand the tech to the CCP. They don’t even need to do that because Microsoft does it for them." Of equal concern, according to experts, is the fact that Microsoft Research Asia serves as a proving ground for some of China’s best tech talent – many of whom later utilize their expertise to start or work at firms that directly compete with US interests with the backing of the Chinese Communist Party. You can read the full article HERE . ### Tech Integrity Project is dedicated to ending this collusion. Tech Integrity Project works to protect the national security and economic competitiveness of the United States by preventing Big Tech companies from aiding America’s adversaries. The organization works to educate policymakers and the public about the problematic business activities of U.S. tech companies in China and other adversarial nations, including capital investment, overseas research, transfer of trade secrets, and selling access to sensitive technologies.