White House Accuses China of 'Industrial-Scale' AI Model Theft

CryptoFrontier

The White House warned on April 23, 2026, that foreign entities, primarily in China, are conducting “industrial-scale” campaigns to copy American artificial intelligence models, according to a memorandum from Michael Kratsios, Assistant to the President for Science and Technology and Director of the Office of Science and Technology Policy. The campaigns employ tens of thousands of proxy accounts and jailbreaking techniques to extract proprietary capabilities and replicate model performance, a method known as distillation attack.

Tactics and Methods

According to President Donald Trump’s administration, the coordinated efforts use “tens of thousands of proxy accounts” to evade detection and exploit jailbreak techniques to systematically extract capabilities. A distillation attack is a method of training a smaller AI model to learn from the outputs of a larger one. Kratsios stated on X (formerly Twitter): “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.”

Historical Precedent

The issue has become a growing concern among U.S. AI companies. In February 2026, Anthropic accused Chinese AI labs DeepSeek, Moonshot, and MiniMax of extracting millions of Claude responses—using roughly 24,000 fraudulent accounts—to train competing systems. Models developed through unauthorized distillation campaigns may not match the full performance of the originals but can appear comparable on select benchmarks at a fraction of the cost.

Security Concerns

The administration warned that distillation attacks could remove security safeguards and other controls designed to keep AI systems “ideologically neutral and truth-seeking.” The Trump administration said federal agencies will work with U.S. AI companies to strengthen protections around frontier models, coordinate with private industry to develop defenses against large-scale distillation campaigns, and explore ways to hold foreign actors accountable.

Administration Response

While the memorandum acknowledged that lawful distillation can help create smaller, more efficient open-source and open-weight models, it stated that unauthorized efforts to copy American AI systems cross the line. “There is nothing innovative about systematically extracting and copying the innovations of American industry,” the memo said.

Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to Disclaimer.

Related Articles

OpenClaw 2026.4.22 Unifies Plugin Lifecycle Across Codex and Pi Harnesses, Reduces Plugin Load Time by Up to 90%

Gate News message, April 23 — OpenClaw, an open-source AI Agent platform, released version 2026.4.22 on April 22, with its biggest change being the alignment of Codex harness and Pi harness lifecycles. Previously, plugins behaved inconsistently across the two harness pathways, with some hooks

GateNews3h ago

Reppo Foundation Secures $20M Commitment from Bolts Capital for AI Training Data Infrastructure

Gate News message, April 23 — The Reppo Foundation announced it has received a strategic funding commitment of $20 million from Bolts Capital to advance protocol development and ecosystem expansion, with a focus on building AI training data infrastructure powered by prediction markets. Reppo's

GateNews3h ago

EU to Draft Guidelines Requiring Google to Provide Third-Party AI Equal Android Access as Gemini

Gate News message, April 23 — The European Union will draft guidelines requiring Google to provide ChatGPT and Claude with the same functionality access on Android as Gemini receives, according to market sources. The move aims to ensure third-party AI applications have equivalent permissions and

GateNews3h ago

Cluster Protocol Raises $5M in Funding, DAO5 Leads Round

Gate News message, April 23 — Cluster Protocol, an AI deep tech and Web3 infrastructure company, announced the completion of a $5 million funding round led by DAO5, with participation from Paper Ventures, JPEG Trading, and Mapleblock Capital. The company's total funding to date reaches $7.75 million

GateNews4h ago

UAE President Discusses AI and Space Opportunities with Musk and Fink

Gate News message, April 23 — UAE President Sheikh Mohamed bin Zayed Al Nahyan held talks with SpaceX founder Elon Musk and BlackRock chairman Larry Fink on potential opportunities in artificial intelligence and the space sector. Musk spoke with the Emirati leader by telephone, according to the UAE

GateNews5h ago

Google Cloud and CVC Partner to Accelerate AI Agent Transformation for Portfolio Companies

Gate News message, April 23 — Google Cloud and private equity giant CVC announced a strategic partnership to accelerate AI agent transformation for CVC and hundreds of companies within its investment portfolio. The collaboration will leverage Google Cloud's artificial intelligence platform and

GateNews5h ago
Comment
0/400
CheckTheBlockchainBeforevip
· 1h ago
Distilling or copying models actually has very low costs; closed-source methods can't prevent it, at most raising the barrier to entry.
View OriginalReply0
GateUser-1bc81bb2vip
· 1h ago
I'm more concerned about the evidence chain: is it specifically weight leakage, API scraping, or the "mirror" on the training data side?
View OriginalReply0
NoSleepBridgevip
· 1h ago
Don't just focus on "copying"; many model capabilities are inherently similar. Data pipelines and productization are the true moats.
View OriginalReply0
DustyAlphavip
· 1h ago
It seems that the AI competition is once again being elevated to a matter of national security narrative.
View OriginalReply0
CapitalFlowInATeacupvip
· 1h ago
AI enters a "arms race" mode, and innovation will accelerate, but ordinary developers may be caught in the middle and feel the most uncomfortable.
View OriginalReply0
0xNapvip
· 2h ago
Could it be an opportunity to promote more localized computing power and chip restrictions? Logically, it makes sense.
View OriginalReply0
GateUser-99725296vip
· 2h ago
The China-U.S. relationship is tightening, and cross-border projects combining Web3 and AI are probably becoming even harder to comply with.
View OriginalReply0
LendingRateAnxietyvip
· 2h ago
The copying of AI models is inevitable; the key is who can deploy computing power, data, and applications faster.
View OriginalReply0