White House warns of 'industrial-scale' efforts in China to rip off U.S. AI tech

```json { "title": "White House Accuses China of Industrial-Scale AI Theft", "metaDescription": "The White House OSTP warns of industrial-scale Chinese campaigns to steal U.S. AI technology using proxy accounts and jailbreaking techniques.", "content": "<h2>White House Warns of Industrial-Scale Chinese Campaigns to Steal U.S. AI Technology</h2><p>The Trump administration formally accused Chinese entities on April 23, 2026 of waging industrial-scale campaigns to steal American artificial intelligence technology, warning that the efforts use tens of thousands of proxy accounts and jailbreaking techniques to extract proprietary information from leading U.S. AI systems. The accusation was detailed in a memo authored by Michael Kratsios, director of the White House Office of Science and Technology Policy (OSTP), and arrives amid an escalating pattern of U.S. government concern over Chinese AI intellectual property theft.</p><h2>What the White House Memo Says</h2><p>The OSTP memo, first reported by the Financial Times, lays out the administration's case that foreign entities — principally based in China — are systematically exploiting a technique known as <strong>distillation</strong>: a process by which a less powerful AI model learns from a more powerful one, gaining capabilities without significant independent research investment. According to the memo, these campaigns allegedly enable bad actors to extract and reproduce the core capabilities of American frontier AI models at scale.</p><p>Kratsios was direct in his assessment. "The US government has information indicating that foreign entities, principally based in China, are engaged in deliberate, industrial-scale campaigns to distil US frontier AI systems," he wrote, as reported by the Financial Times. In a post on X the same morning, he added: "The U.S. has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI. We will be taking action to protect American innovation."</p><p>Beyond simple capability theft, the memo raised a more pointed concern: that the distillation process is being weaponized to strip safety features out of the resulting models. "These distillation campaigns also allow those actors to deliberately strip security protocols from the resulting models and undo mechanisms that ensure those AI models are ideologically neutral and truth-seeking," Kratsios wrote.</p><p>The White House said it will explore a range of measures to hold foreign actors accountable and will share information with U.S. AI companies about the adversarial campaigns. The Chinese Embassy in Washington did not immediately respond when asked for comment on the memo, according to CNBC.</p><h2>Senate Hearing Puts IP Theft Numbers on the Record</h2><p>The OSTP memo landed just one day after the Senate Judiciary Committee held a hearing titled <em>Stealth Stealing: China's Ongoing Theft of U.S. Innovation</em> on April 22, 2026 — a pairing that underscores how urgently the issue is being treated across branches of the U.S. government.</p><p>In his statement to the committee, Senator Chuck Grassley cited estimates that China steals between $400 billion and $600 billion of U.S. intellectual property each year — a figure that works out to approximately $5,000 per American taxpayer. The hearing also featured testimony from Tom Lyons, a former CIA officer with more than 20 years of experience in the U.S. government and private sector on Chinese economic espionage.</p><p>Grassley's statement included a striking case study: a Chinese national was found guilty of stealing trade secrets from Google and, between May 2022 and April 2023, uploaded over 2,000 pages of confidential information containing AI trade secrets to his Google Cloud account before returning to China to found his own AI and machine learning company.</p><h2>How the AI Industry Got Here: DeepSeek and the Distillation Debate</h2><p>Allegations of Chinese distillation attacks against American firms did not begin with this week's memo. They surfaced in earnest following Chinese company DeepSeek's release of its R1 reasoning model in January 2025. At the time, OpenAI and Microsoft alleged that DeepSeek had been trained on outputs from ChatGPT — a claim that raised immediate questions about how a Chinese lab with comparatively limited access to advanced semiconductors had developed such a capable model so quickly.</p><p>OpenAI formalized its allegations in February 2026, publicly releasing a memo to the U.S. Congress's China Select Committee on February 12 alleging that DeepSeek had stolen its intellectual property to fuel its own models, according to the Foundation for Defense of Democracies (FDD).</p><p>By April 2026, the industry response had taken a more coordinated shape. On April 6, 2026, OpenAI, Anthropic, and Alphabet announced plans to share information with one another related to Chinese-linked industrial espionage through the Frontier Model Forum, according to the FDD. The move marked an unusual moment of collaboration among direct competitors, underscoring the shared threat perception within the American AI sector.</p><p>The Trump administration had previously proposed standing up an AI-Information Sharing and Analysis Center (AI-ISAC) led by the Department of Homeland Security as part of its AI Action Plan released in July 2025. However, as of February 2026, there was no timeline for establishing it, according to the FDD.</p><h2>Expert Reactions</h2><p>The verified statements from officials and analysts paint a picture of a threat that goes beyond conventional corporate espionage. Kratsios, in his CNBC-reported remarks, was unambiguous about how the administration views the scale of what is alleged: "Industrial distillation activities that aim to systematically undermine American research and development and access proprietary information, however, are unacceptable." He added: "There is nothing innovative about systematically extracting and copying the innovations of American industry."</p><p>At the Senate Judiciary Committee hearing, former CIA officer Tom Lyons framed the competitive dynamic in stark terms: "American firms are not competing against Chinese rivals in any normal sense."</p><p>Chris McGuire, whose title was not specified in available sourcing, offered a more technical characterization of the strategic logic behind the attacks: "Chinese AI firms are relying on distillation attacks to offset deficits in AI computing power and illicitly reproduce the core capabilities of US models."</p><h2>What's Next: Accountability Measures and a Looming Summit</h2><p>The timing of the OSTP memo carries diplomatic weight. According to Fox Business, the White House accusation was made just weeks before a planned summit between President Donald Trump and Chinese President Xi Jinping, originally scheduled for late March 2026 but postponed to May 14. How — or whether — the AI theft allegations will feature in those discussions has not been publicly specified.</p><p>What the White House has committed to is sharing threat intelligence with U.S. AI companies about the adversarial campaigns and exploring accountability measures for foreign actors. The specific nature of those measures has not been detailed in publicly available information as of April 23, 2026.</p><p>The Frontier Model Forum collaboration between OpenAI, Anthropic, and Alphabet represents the most concrete industry-level response to date, though it too is in early stages. The proposed AI-ISAC under the Department of Homeland Security remains without an implementation timeline.</p><p>For now, the combination of the OSTP memo, the Senate Judiciary hearing, and the industry's own information-sharing efforts signals that U.S. institutions — governmental and corporate alike — are treating industrial-scale AI distillation as a systemic national security issue rather than a series of isolated incidents.</p><p>For more tech news, visit our <a href=\"/news\">news section</a>.</p>", "excerpt": "The Trump administration formally accused Chinese entities on April 23, 2026 of running industrial-scale campaigns to steal U.S. artificial intelligence technology through distillation attacks using tens of thousands of proxy accounts. The accusation, detailed in a memo by White House OSTP Director Michael Kratsios, arrived one day after a Senate Judiciary Committee hearing on Chinese IP theft and weeks before a planned Trump-Xi summit. U.S. AI firms OpenAI, Anthropic, and Alphabet have already begun coordinating through the Frontier Model Forum in response to the shared threat.", "keywords": ["China AI theft", "industrial-scale distillation attacks", "White House OSTP", "Michael Kratsios", "AI intellectual property"], "slug": "white-house-accuses-china-industrial-scale-ai-theft" } ```