Yesterday, March 25, U.S. Senator John Barrasso (R-WY), ranking member of the Energy and Natural Resources (ENR) Committee, sent a letter to Secretary of Energy, Jennifer Granholm, outlining his concern about the ability of the Department of Energy (DOE) to protect its cutting-edge research and development (R&D) into artificial intelligence (AI) from China.
WASHINGTON, DC - AUGUST 05: U.S. Sen. John Barrasso (R-WY) speaks at a press conference at the U.S. Capitol on August 05, 2022 in Washington, DC. The group of Republican Senators held the press conference to speak out against the Democrats' tax and spending policies. (Photo by Kevin Dietsch/Getty Images)
In the letter, ranking member Barrasso highlights the Chinese Communist Party’s (CCP) attempts to target pioneering AI R&D, emphasizing the necessary steps DOE must take as its leadership in federal AI R&D expands. These actions include remediating documented counterintelligence deficiencies, discontinuing the active sharing of R&D with official members of the CCP, and fully funding the DOE Office of the Inspector General.
“In the face of mounting threats posed by the CCP’s pursuit of technological dominance in AI, DOE’s active engagement with CCP-affiliated individuals and organizations is recklessly foolish. This engagement must cease immediately,” Barrasso wrote.
Read the full letter here and below:
Dear Secretary Granholm,
I am writing to express my deep concern about the ability of the Department of Energy’s (DOE) research security apparatus to resist threats from the Chinese Communist Party (CCP). This concern is especially heightened due to the continuing expansion of DOE’s leadership in federal research and development (R&D) for artificial intelligence (AI), an area the CCP actively seeks to dominate. In order to counter the CCP’s relentless targeting of pioneering AI R&D within DOE, the Department must accomplish three critical tasks:
remediate its chronic counterintelligence shortcomings and fully inform Congress of its progress in such efforts;
cease active R&D sharing with official members of the CCP, as well as individuals and organizations under the control of the CCP; and
fully fund the DOE Office of the Inspector General (DOE OIG)
It is essential for DOE to understand that broad congressional support for its expanded AI leadership hinges upon addressing these actions promptly and effectively.
Addressing DOE’s Chronic Counterintelligence Shortcomings
As you are aware, I, along with several members of the U.S. Senate Select Committee on Intelligence, sent you a letter (attached) on November 21, 2023, detailing significant concerns about the state of counterintelligence within DOE and its National Laboratories. I remain particularly troubled by the abrupt reassignment of DOE’s Director of Intelligence and Counterintelligence. Likewise, I continue to be troubled by a classified report by an outside contractor highlighting systemic counterintelligence flaws within DOE. By failing to provide adequate responses to the questions in our November 21st letter, DOE has dramatically slowed our efforts to gather information and fulfill our oversight duties. This is wholly unacceptable.
The recent indictment of Linwei Ding, a Chinese national and a former Google employee charged with stealing over 500 confidential files related to Google’s AI efforts, underscores the clear and present danger posed by the CCP’s efforts to obtain American AI secrets. This case reinforces the critical need for DOE’s enhanced protection of AI technologies to safeguard national security interests.
Decoupling from CCP Influence and Engagement
On March 4, 2024, I sent a letter (attached) expressing concerns about the DOE’s offer of taxpayer-funded research to official members of the CCP, including at least one agent of the United Front Work Department. In response, DOE made the outlandish claim to news outlets that the sharing of research with these compromised individuals will not only continue but will enhance America’s national security. According to a DOE spokesperson, “…department officials have and will continue to strategically engage with partners and competitors from across the globe to protect and promote American innovation…and strengthen our national security.”
This eagerness to engage with the PRC via scientific collaboration clearly contradicts a recent report released by the National Science Foundation titled Safeguarding the Research Enterprise. A key finding of the report states, “recent efforts of the [PRC] to preferentially direct fundamental research toward military needs, and its decision to restrict the flow of information out of the country, may severely limit the benefits of collaborations with research organizations within the PRC.”
In the face of mounting threats posed by the CCP’s pursuit of technological dominance in AI, DOE’s active engagement with CCP-affiliated individuals and organizations is recklessly foolish. This engagement must cease immediately. It is a fact that AI companies in the People’s Republic of China (PRC) overcome innovative shortfalls by using open-source American models to fill innovation gaps and stay competitive with the U.S. in the global AI race. Therefore, DOE must implement a robust vetting process for those wishing to access and work with DOE-created AI foundation models. It is crucial that those friendly to the CCP do not gain access.
Sufficiently Funding the DOE OIG
To ensure proper oversight of DOE’s efforts to mitigate the CCP’s theft of pioneering AI R&D, DOE must fully support the DOE IG’s request to close the funding gap for the OIG. According to the IG, “With the current level of proposed [FY 2025] funding… the OIG’s oversight would be a fraction of what it should be and would not include any oversight of many key areas…For these reasons, I must conclude that I am substantially inhibited from performing the duties of my office.”
Among these duties is oversight of “high risk areas,” including “[t]heft of intellectual property and research impacting national security,” as well as the “[n]ew and expanding artificial intelligence program.” According to the IG:
The OIG is responsible for the audit and evaluation of the Department’s unclassified systems. The Department has experienced substantial problems with cybersecurity…[I]t will become increasingly important to secure its systems from vulnerabilities that could result in the loss of billions of dollars’ worth of innovative or sensitive technologies developed using taxpayer dollars.
As U.S. officials, including FBI Director Christopher Wray, express grave concerns over PRC hacking and insider theft of AI secrets, it is essential that the DOE OIG has the resources to oversee the protection of DOE’s pioneering AI R&D, particularly from adversarial nations like the PRC. The security of our nation's technological advancements and national interests must remain paramount.
I ask that you respond to this letter by April 8, 2024, detailing a plan of action for how DOE plans to address each of the three issues highlighted above. Additionally, you have not responded to me directly regarding my attached March 4, 2024 letter. I ask that you respond by April 1, 2024.
Sincerely,
Comments