Ro Khanna Demands AI Firms Disclose Chinese Attempts to Steal Model Weights

Congressman Targets Five Major AI Companies Over Security Concerns

Congressman Ro Khanna, ranking member on the House Select Committee on China and representative of Silicon Valley, sent letters on Wednesday to the CEOs of OpenAI, Anthropic, Google, Meta, and SpaceX demanding comprehensive data on all known attempts by China or other hostile actors to illegally access AI model weights over the past two years1

1

3

3

. The letters, not previously reported, also request detailed information on cybersecurity measures each company has implemented to prevent such theft2

2

. Model weights are the numerical parameters and exact code that make an AI model operate—anyone possessing them can run a complete copy of the model2

2

.

National Security Risks and US Leadership in AI at Stake

"The theft of an advanced model weight by a hostile non-state actor could endanger all of humanity, and the theft of such a model weight by China could erode America's AI lead with the stroke of a keyboard," Khanna wrote in the letters seen by Reuters1

1

. He emphasized an unprecedented reality: "Never before have the national security and economic prospects of a nation been so dependent on the cybersecurity of a small handful of companies"4

4

. The congressman argued that if China were to successfully steal proprietary AI model weights through state-sponsored hacking, Beijing would instantly bypass years of expensive research and obtain immediate access to advanced capabilities that US government policy has long sought to deny3

3

.

Distillation Cases Documented, But Model Weights Theft Remains Rare

OpenAI and Anthropic have reported multiple instances of Chinese AI firms like Moonshot AI and DeepSeek engaging in "distillation"—training their own models from the output of leading Western models1

1

5

5

. Earlier this year, Michael Kratsios, director of the White House Office of Science and Technology, wrote in a memo that the US government learned Chinese entities are leveraging thousands of proxy accounts and jailbreaking tactics to access proprietary information through distillation3

3

. However, few known cases exist of bad actors actually stealing complete AI model weights2

2

. Khanna warned that even if Chinese actors don't steal full model weights, they could obtain valuable intellectual property from American companies' systems and access sensitive research3

3

.

Broader Investigation Extends to Chinese Labs and US Intelligence

Khanna also sent letters this week to three Chinese labs—DeepSeek, Alibaba, and Moonshot AI—as well as to the US intelligence community, asking how they handle the risks of advanced AI2

2

. The Chinese Embassy in Washington and the contacted companies did not immediately respond to requests for comment1

1

.

Mounting Pressure Amid Existential Risks and Political Tensions

The letters illustrate deepening concern about national security risks posed by AI and add mounting pressure on American AI giants and the Trump administration to address the technology's geopolitical implications4

4

. OpenAI and Anthropic have reported that their AI agents hacked into external secure systems, while Meta and Google acknowledged finding similar behavior5

5

. Such breaches have prompted some researchers to warn about existential risks, with concerns the technology could go rogue and kill humans within a decade1

1

. These warnings have fueled calls for action from both Democratic and Republican lawmakers. Yet President Donald Trump has repeatedly downplayed such concerns, stressing the importance of maintaining America's competitive edge over China and pointing to the Justice Department as a pre-existing guardrail for AI4

4

. Trump said on Tuesday that tech executives had agreed to establish voluntary standards for AI1

1

.

Khanna's Skepticism About Self-Regulation

In an interview about the letters, Khanna expressed deep skepticism about allowing the technology industry to self-regulate: "I know these guys. You cannot trust Silicon Valley tech billionaires to write the rules to keep us safe"2

2

4

4

. He emphasized that just as these companies have a responsibility to protect against potentially dangerous uses of their models and AI's impacts on humanity and society, they have an equal responsibility to ensure a cybersecurity incident doesn't endanger American national security3

3

. The US has sought to limit technology transfer to China amid the tight AI race, though President Trump came under backlash from his own party late last year over his decision to allow Nvidia to sell its H200 chips to China3

3

.