Senator Banks Raises Concerns Over Chinese Espionage Targeting U.S. Artificial Intelligence Sector

WASHINGTON, D.C. - Yesterday, Senator Jim Banks (R-Ind.) and Senate Judiciary Chairman Chuck Grassley (R-Iowa) sent a letter to nine individual artificial intelligence companies citing growing concerns that the People's Republic of China (PRC) is actively targeting America's AI sector through espionage and other security threats. The letter highlights mounting evidence that the Chinese Communist Party (CCP) is leveraging state-backed programs, corporate infiltration, and coercive tactics to access sensitive AI technologies critical to U.S. national security and economic leadership. The letter also asks AI companies to describe how they detect and guard against PRC espionage, how they manage insider threats, whether they are capable of preventing PRC actors from stealing their models, and whether they have policies to notify the U.S. government if they detect security threats.

The 9 letters were sent to the following:

• Sam Altman, Chief Executive Officer, OpenAI
• Dario Amodei, Chief Executive Officer, Anthropic
• Sundar Pichai, Chief Executive Officer, Google LLC
• Elon Musk, Chief Executive Officer, x.AI Corp.
• Mark Zuckerberg, Chief Executive Officer, Meta Platforms, Inc.
• Satya Nadella, Chief Executive Officer, Microsoft
• Andy Jassy, Chief Executive Officer, Amazon
• Ilya Sutskever, Chief Executive Officer, Safe Superintelligence Inc.
• Mira Murati, Chief Executive Officer, Thinking Machines Lab

Read more about the letter here.

Read the nine company specific letters here or see the standard version of the letter below:

Dear XXX:

We are writing with concerns regarding espionage and related security threats to artificial intelligence (AI) posed by the People's Republic of China (PRC). As you are aware, the PRC will do whatever it takes to overtake U.S. AI leadership, including by means of corporate espionage.

Specifically, the Chinese Communist Party (CCP) has an extensive track record of conducting espionage on U.S. companies in critical sectors. For instance, earlier this year, a jury convicted a Chinese national and former Google employee of stealing proprietary information about Google's AI development. The DOJ found that he was motivated by Chinese national policies encouraging the development of the Chinese AI industry. A later superseding indictment describes how PRC-sponsored programs incentivize researchers outside China to transmit their knowledge and research back to China in exchange for salaries, research funds, lab space, or other incentives.

Moreover, an April 2025 report revealed several vulnerabilities of frontier AI research to espionage by the Chinese government. The report declared that the Chinese Ministry of State Security (MSS) targets ethnic Chinese as a matter of national policy. It further noted that the CCP engages in intimidation and pressure campaigns against Chinese citizens and ethnic Chinese individuals abroad who work on frontier AI. The report also highlighted that those with familial or financial ties to the Chinese mainland are particularly susceptible to intimidation and pressure from the Chinese government. Protecting AI technology from PRC espionage is of paramount importance, especially as these systems become more powerful.

The threat of espionage is especially concerning in the field of AI, given unique considerations around model weights-the "learnable parameters that encode the core intelligence of an AI." As you are aware, model weights can be digitally transferred to the PRC. If this occurs, it is less like the PRC stealing a blueprint and more like the PRC stealing a finished AI product. As a result, model weights are an especially valuable target for the CCP, and we should expect Chinese hackers to be highly motivated to find ways around access controls and other defenses.

With this context in mind, we request written responses to the following questions by May 26, 2026:

1. What is XXX's current approach to personnel vetting, insider threat detection, and ongoing monitoring for privileged access roles?

1. Based on XXX's interactions with the broader frontier AI ecosystem (including public information and voluntary information-sharing), what common insider-threat vulnerabilities do you believe are shared among frontier AI companies? What mitigation strategies are most important for companies to implement?

1. Do you currently believe your company can robustly secure model weights and other sensitive insights from PRC threat actors? If not, what new techniques, policy processes, or other activities are needed to achieve this level of security?

1. How many people have direct or indirect privileged access to model weights or weight-adjacent assets, and how has that number changed over time?

1. How many PRC nationals does XXX employ? How many have direct or indirect privileged access to weights or weight-adjacent assets, and how has that number changed over time?

1. Which roles or levels of access, if any, does XXX consider so sensitive that PRC nationals are not permitted to hold them? Which roles or levels of access are considered most sensitive, and for these, what personal screening, monitoring, or other insider threat mitigation measures are taken?

1. In the event XXX detects, suspects, or reasonably assesses that model weights or other sensitive information was accessed, copied, or exfiltrated by a PRC or foreign-linked actor, what is XXX's notification policy to the U.S. government? If there is not a notification policy, please state so and explain why.

1. In the event that XXX detects, suspects, or reasonably assesses that an XXX model or agent attempted to exfiltrate model weights, communicate sensitive information to the PRC, or otherwise acted contrary to the operator's intent in a way that threatened U.S. national security, what is XXX's notification policy to the U.S. government? If there is not a notification policy, please state so and explain why.

1. What support or engagement from Congress or the U.S. government would be useful in securing AI technology, trade secrets, and research from the PRC?

Thank you for your prompt review and response. This century may be defined by advanced artificial intelligence. Your work to advance U.S. AI leadership and protect this technology from the PRC is essential for continued American leadership in this area and national security. We look forward to your responses.

###