To limit Chinese influence on commercial tech partners, Pentagon plans big changes
New rules aimed at exposing foreign influence over companies the Pentagon does business with are at odds with Defense Department efforts to work with more startups, but there could be a way to do both—if the DOD changes the way it approaches risk, security experts say.
U.S. efforts to secure classified weapons information have worked so well that China is devoting more and more time to scouring unclassified information to gather intelligence on new potential capabilities, Matthew Redding, the assistant director for industrial security at the Defense Counterintelligence and Security Agency, said last week at a Potomac Officer’s Club event in Virginia.
“They’re moving away from [searching for intelligence] behind the classified castle walls, which are very secure under the National Industrial Security Program. While the classified world is very secure, what about that unclassified storefront? What about that unclassified research outside of your classified research areas, right?” he said. His office will collect information on defense contractors in order to evaluate whether some issue related to foreign influence might prohibit that contractor from doing classified work.
The 2020 National Defense Authorization Act mandated a lot more vetting and scrutiny of defense contractors to detect possible Chinese influence. Any Defense Department contract over $5 million will be subject to increased vetting by Redding’s agency, once the policy is reviewed thoroughly.
That means that any company with even a small contract to do classified work will need to submit a form essentially outlining any possible foreign influence or investment. “I’ve been in the federal government for 35 years. This is the largest defense acquisition reform I have seen in 35 years,” Redding said.
A delicate balance
Declaring possible foreign investment exposure is pretty simple for an established defense contractor. But for the new class of tech companies that also sell to consumer markets, it’s much more complicated, because Chinese investment in Silicon Valley startups is high and Chinese researchers are very active in fields like AI and quantum computing.
That means that the Pentagon runs the risk of pushing out exactly the sorts of companies and new tech players it wants to engage. But a little-known DARPA program shows a potential way forward.
DARPA’s Countering Foreign Influence Program takes a data-centered approach to illuminate not just the existence of possible Chinese influence in a given company’s investor board, products, or research, but also the level of security risk posed by that investor, researcher or product. DARPA applies it to its own process of awarding research grants in several high-tech areas.
Scott Myers, who leads the DARPA program, said research partners in academia resisted the idea when DARPA first implemented it in 2022, calling it McCarthyism. “There was a massive outcry even from the scientists in my own building, saying we’re having a chilling effect on science, we’re having a chilling effect on research.”
But Myers, who also spoke at the Potomac Officer Club conference, said research partners warmed to the idea when they learned the model wasn’t intended to prohibit them from working with, say, top researchers who happened to be foreign. Instead, it aims to enable them to do that safely—even if that means creating special boundaries or rules on a case-by-case basis.
“We believe that we can get to ‘yes’ even if the risks are very, very high. We do everything in our power to mitigate those risks. The way we do that is we make sure that our partners have skin in the game.”
In one example he cited, a university that was partnering with DARPA on a program wanted the pre-eminent expert in the field on their team, but that person had studied at the Chinese Academy of Sciences, where all research findings belong to the Chinese government. The PRC also uses a variety of tactics to coerce Chinese researchers living abroad to share information and secrets with them.
The university came back with a proposal to allow the researcher to work on the program by isolating them—physically and in terms of communication and information—from other parts of the research, he said.
“Their proposal was that ‘We can, we can do this. The individual will be escorted into a room where they’re by themselves. They will be on their own network, completely air-gapped from everyone else. They will only receive information for their portion of the work. Once they complete their work for the day, they leave for the day they will leave out a door that nobody else has walked out of. They will get no results from any other parts of the program. They will get no final results until the final paper is published, thus preventing them from getting a patent for the information in foreign country’,” Redding said.
But DARPA’s solution is not the only option. Lisa Sanders, the director of science and technology at U.S. Special Operations Command, suggested another possibility, based on SOCOM’s experience.
Because special operators spend a lot of time working with partner militaries that don’t have access to classified information, SOCOM has found ways to share important information in unclassified environments. She described a 2022 workshop led by SOCOM with members of the Norwegian military who had unique data and insight on the way the electromagnetic spectrum works in the Arctic Circle—information that’s relevant to operating in other places where Russia conducts a lot of information warfare.
So though the military treats that topic as one of the most sensitive, SOCOM led a three-day forum she described as a “completely open, notional, unclassified conversation” —and it produced useful results.
That approach opens the possibility of the information getting to adversaries, who could develop the same capability later. But, Sanders said, in some instances, it’s better to risk Russia, China or another adversary stealing a piece of research or breakthrough rather than not getting it to operators quickly. Then, the trick is staying ahead of the innovation that comes after.
Comments are closed.