Beijing's Spy Chief Warns AI Could Undermine Communist Party Control
Days after Washington accused six Chinese firms of mass-scale AI model theft, China's state security chief raises his own alarm about AI risks to Party rule.
Part of a larger narrative
Trump's Ireland Visit
- Burnham Tells UN Russia Uses AI-Boosted Disinformation, Unveils UK Defence Centre
- Alibaba unveils Zhenwu V900 AI chip, plans 5-10 trillion parameter model
- OpenAI, Anthropic CEOs tell UN Security Council AI could risk 'humanity as a whole'
- State Department tells diplomats to call AI 'super intelligence' after Trump's UN rename
- US proposes AI incident "hotline" with China ahead of Trump-Xi summit
What to know
- US agencies (FBI, NSA, CISA) accuse six Chinese AI firms — DeepSeek, Alibaba, Moonshot AI, MiniMax, StepFun, Z.AI — of systematically distilling American frontier models since 2024, calling it the core of China's AI strategy.
- China's foreign ministry denies wrongdoing and frames its AI progress as self-reliance, while separately its top state security official warns AI itself could threaten Communist Party control.
- The dispute unfolds ahead of a planned September 24 Trump-Xi meeting, with Trump publicly dismissing Anthropic CEO Dario Amodei's AI-safety warnings and stressing the US must not lose the AI race to China.
- Online reaction is largely skeptical of the US framing, with commenters calling the accusation hypocritical given American AI firms' own reliance on scraped global data and questioning why a terms-of-service breach is treated as a national security matter.
The dispute Whether Chinese distillation of US models constitutes a genuine national-security threat or is hypocritical grandstanding given American firms' own unlicensed data scraping. · positions read across 37 posts and comments
The US accusation is hypocritical since American AI firms themselves scraped the world's data without consent to train their models.
-
“And American models copied the whole world's data illegally to train their shitty models. If you want to enforce the law, start with your own American companies who stole every book, form of art and everything else out there.”
nixCraft · Mastodon ↗
Treating a terms-of-service violation as a national security incident is an overreaction that trivializes the term.
-
“the master criminal they caught... violated the Terms of Service. the checkbox. "I have read and agree." nobody has ever read it. it is now a national security incident”
k3ym0 · Mastodon ↗
“Working with Beijing would be a mistake. The Chinese government's true intentions are clear.”
WSJ Opinion, Editorial board · Wall Street Journal ↗ · Sep 13
Mao Ning Chinese Foreign Ministry spokespersonChen Yixin Head of China's Ministry of State Security
Donald Trump US President
Dario Amodei Anthropic CEO
Michael Kratsios Director, White House Office of Science and Technology PolicyDeepSeek, Alibaba, Moonshot AI, MiniMax, StepFun, Z.AI Accused Chinese AI developers
How it unfolded 6 developments, newest first · click a bar or a number to jump articlesvideospostscomments
-
6
Commentary frames Chen Yixin as China's own AI doomsayer
The Register characterized Chen Yixin's warnings as China's version of AI-risk alarmism, noting Beijing's response is to tighten technological sovereignty and impose broad regulations even as it publicly downplays US theft accusations.
“Beijing's response is to 'firmly grasp technological sovereignty' and broad regulations…”
— The Register -
first by The Register, 9d ago
-
🇨🇳China’s intelligence chief says AI could UNDERMINE the Communist Party’s power. State Security official Chen Yixin wants tighter party control and tougher government oversight to prevent AI from becoming an active threat to China. Meanwhile, the rest of the world is worrying
2 more of the top 3 · 14 posts in this stretch
-
R
"China’s minister for State Security has decided AI might be bad for the nation’s ruling Communist Party. Party secretary and minister Chen Yixin’s views appeared in China Cyberspace Magazine, the flagship publication of China’s Cyberspace Administration (and which readers may recall once carried a piece by Elon Musk). In Chen’s view, “the field…
-
But that's exactly why I said relatively. Astra is at the very top end, and the only genuinely comparable model is Fable/mythos, which is in the same general price class. Saying “we managed to spend $100k on Astra” shows that heavy frontier model usage can be expensive in absolute terms. It doesn't show that a much cheaper model could have handled…
-
-
5
Trump criticizes Anthropic's Amodei over AI warnings
In a CBS News segment, President Trump criticized Anthropic CEO Dario Amodei's warnings about AI risk and reiterated that the US must maintain its lead over China in the AI race.
-
T
« #China’s #spy agency has warned that #AI could pose a #risk to the country’s political and social #security , signalling growing concern in Beijing that rapid advances in the #technology could be exploited against the country. ft.trib.al/l7hOWIG Hostile forces using AI for ‘c...
1 more of the top 2 · 2 posts in this stretch
-
The biggest barrier to a US-China AI safety treaty are that the two countries are led by 80- and 73-year-old narcissists who regularly wipe “disloyal” experts on the tech from their govt and see everything as zero-sum. Actually very dangerous.
-
-
4
China's state security chief Chen Yixin flags AI as a risk to Party control
Reporting described China's Ministry of State Security head Chen Yixin warning that AI could pose a risk to the country's political and social security, urging tighter Party control and government oversight, part of a broader Chinese AI safety posture emphasizing 'technological sovereignty'.
“China's intelligence chief says AI could UNDERMINE the Communist Party's power.”
— @MarioNawfal -
first by SCMP China, 10d ago · also Bloomberg Tech, FT, Paper Tiger, South China Morning Post, SCMP
5 more headlines
- China Spy Chief Warns of AI Risks as US Tech Leaders Urge Brakes Bloomberg Tech · 10d ago
- China’s spy agency warns of AI risk to national security Financial Times · 9d ago
- China Spy Chief Warns of AI Risks as Anthropic CEO Urges Brakes Bloomberg.com · 9d ago
- China's AI Reckoning Paper Tiger · 9d ago
- China's intelligence chief warns of risks from AI in hands of ‘hostile forces’ South China Morning Post · 9d ago
-
first by Reuters, 9d ago · also Investing.com News, Modern Diplomacy
2 more headlines
- Explainer-How China is preparing for the risk of AI escaping human control Investing.com News · 9d ago
- China's AI Safety Strategy: How Beijing Is Preparing for Loss of Human Control Modern Diplomacy · 9d ago
-
The Chinese government has now clearly and unmistakably rejected serious collaboration with the United States on AI safety. This was a predictable development. It is also a clarifying moment for the importance of U.S. export controls on AI chips and other technologies, which are now indisputably a ...
2 more of the top 3 · 8 posts in this stretch
-
M
China joins the chorus. China’s spy agency warns of AI risk to national security https://www. ft.com/content/8715d1c6-054d-4 eab-bcad-147acebfd2a9?syn-25a6b1a6=1
-
China joins the chorus. — China's spy agency warns of AI risk to national security —
-
- 4 days quiet
-
3
Online commentary mocks and disputes the US advisory
Commenters on Mastodon and Hacker News called the accusation hypocritical given American AI firms' own scraping of global data, and derided the underlying wrongdoing as merely a terms-of-service violation being treated as a national security matter.
“the master criminal they caught... violated the Terms of Service. the checkbox. "I have read and agree." nobody has ever read it. it is now a national security incident…”
— k3ym0 -
6 outlets China hits back at US claims of ‘malicious’ AI distillation ahead of planned Trump-Xi talks
first by Mercury News, 14d ago · also AP via Feedx, Philadelphia Inquirer, Boston Herald, The Mercury News, Associated Press
-
M
Sure, Jan... Six Chinese AI firms accused of aggressively copying US frontier models https:// arstechnica.com/tech-policy/20 26/09/six-chinese-ai-firms-accused-of-aggressively-copying-us-frontier-models/ # China # AI # Models # Tech
1 more of the top 2 · 3 posts in this stretch
-
Oh no! American AI companies don't like industrial-scale knowledge distillation campaigns! 🎻 https://www. cisa.gov/news-events/cybersecu rity-advisories/aa26-251a
-
-
2
China's foreign ministry rebuffs the distillation accusation
Ministry of Foreign Affairs spokesperson Mao Ning said she had not seen the report but that China's AI development is a result of self-reliance, and urged the US to cooperate rather than make 'groundless accusations'.
“We always believe that all parties should strengthen cooperation to foster open, inclusive, beneficial and ethical development of AI for the well-being of all humanity.”
— Mao Ning, Chinese Foreign Ministry spokesperson -
first by Hong Kong Free Press HKFP, 13d ago · also Breitbart, Channel News Asia
1 more headline
- China Pushes Back Against American Allegations of ‘Malicious’ AI Theft Breitbart · 13d ago
-
first by WSJ Opinion, 9d ago
-
R
"Chinese artificial intelligence firms are committing “industrial-scale” theft of their American counterparts’ trade secrets to save money and bring the coveted technology to market faster, US federal agencies alleged Tuesday. Major Chinese tech companies are using “distillation,” which involves training lesser AI models on more advanced ones, to…
2 more of the top 3 · 4 posts in this stretch
-
U
https://www. europesays.com/us/1052440/ US claims Chinese AI firms are carrying out ‘industrial-scale’ theft of trade secrets # ai # ArtificialIntelligence # Technology # UnitedStates # UnitedStates # US
-
J
CISA Says DeepSeek, Alibaba and Others Extracted Billions of Tokens From US Frontier AI Systems. https://www. govinfosecurity.com/us-warns-c hinese-ai-firms-are-illicitly-distilling-models-a-32773
-
-
1
US agencies accuse six Chinese AI firms of mass model distillation
A joint advisory from the FBI, NSA and CISA said Chinese companies DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun and Z.AI had since 2024 systematically distilled American frontier models — Anthropic's Claude, OpenAI's ChatGPT, Google's Gemini and xAI's Grok — calling the practice the 'core' of China's AI strategy.
“The sheer scale of these campaigns and their sophistication indicate that distillation is not a supplement to these companies' AI model development, but the critical core of it.”
— FBI/NSA/CISA joint advisory -
first by Daily Maverick, 14d ago · also Reuters, SCMP China, SCMP, Straits Times
-
first by Investing.com News, 15d ago · also Hong Kong Free Press HKFP, CyberInsider, The Information, Channel News Asia
4 more headlines
- US Accuses Six Chinese AI Firms of Industrial-Scale Model Distillation Techstrong.ai · 15d ago
- Washington accuses Chinese AI firms of ‘systematic’ theft Hong Kong Free Press HKFP · 15d ago
- Chinese AI firms use industrial-scale distillation to copy US models CyberInsider · 15d ago
- U.S. Accuses DeepSeek, Alibaba and Other Chinese AI Firms of Large-Scale Distillation The Information · 15d ago
-
R
In other news, 🇺🇸 🇨🇳 U.S. cyber and law enforcement officials on Tuesday accused Chinese # AI companies of "aggressive, industrial-scale distillation activities," according to a statement from three U.S. security agencies. Distillation is the process of training smaller AI models using output from larger, more expensive ones as…
2 more of the top 3 · 6 posts in this stretch
-
Major cybersecurity advisory on adversarial distillation with extensive and detailed attribution of PRC AI developers. With these facts established, the U.S. government should permit U.S. companies to work together to detect and disrupt these campaigns, as the report recommends.
-
V
Hah, US AI companies are again accusing their Chinese counterparts of stealing "intellectual property". Obviously, only US tech companies are allowed to steal IP at a massive scale. Hypocrisy is just too mild a word for it ... https:// edition.cnn.com/2026/09/08/pol itics/us-accuses-china-of-stealing-ai-technology # AI # Tech # US # China
-
Also covered reported alongside — the timeline has no entry for these yet
-
first by Yahoo News Singapore, 14d ago · also CNN, KVIA-TV
1 more headline
-
first by Al Jazeera, 14d ago · also Reuters
-
first by CISA, 15d ago · also Reuters
and 16 smaller pieces
What people are saying 12 voices from 3 sites · best of 37 · verbatim
- Will the distillation dispute come up when Trump meets Xi Jinping on September 24?
- Does Chen Yixin's warning about AI threatening Party control reflect a genuine shift in Beijing's internal AI policy or just rhetoric?
- Sep 15
-
What makes you describe chatgpt as just 'good enough', Astra is out and out better for anything substantive and delivers far, far higher usage from the $200 plan.
-
T
China's spy chief identified Anthropic’s Claude Mythos and OpenAI’s GPT-5.5-Cyber as signs of what he called a “disruptive upgrade” in cyber capabilities. https:// therecord.media/china-spy-chie f-warns-of-us-ai-models
-
It isn't. I've looked at the API rates. It's cheaper to write the cache, but cache read being charged at 20% makes it a net loss for agentic coding in a frontier to frontier comparison.
-
O
「最新のAI終焉論者は中国の情報機関トップだ /北京の対応は「技術主権をしっかりと掌握する」ことと、広範な規制を行うことだ。 」： # TheRegister 「中国国家安全部長官は、人工知能（AI）は中国共産党にとって有害となる可能性があると判断した。 陳一新党書記兼大臣の見解は、中国国家インターネット情報弁公室の主要刊行物である中国サイバースペースマガジンに掲載された（読者の中には、かつてイーロン・マスク氏の記事を掲載したことを覚えている人もいるかもしれない）。…
-
This would be one of the reasons to use truly local models - better understanding of the nuances of the language. We have 24 official languages in the EU and about 60 protected dialects. Indians are using their local AI neoclouds to improve support for over 40 local languages in India - we should follow.
-
Also towards big tech, US over EU, and Grok is republican leaning. But does political ideology matter when I want to have some privacy by running the model in my home as a tool? Qwen and Deepseek are a lot better than gpt-oss or Gemma when it comes to coding or logic.
-
Not necessarily, fine tuned to specific tasks models can perform better while being drastically cheaper to run. E.g. Airbus are fine tuning Mistral models on their maintenance manuals which are way too much content to just slap on top of an existing proprietary model with a RAG or something, and they're reporting better results for much less…
-
If you want big open weighted model internally you need infrastructure and there aren't that many places in Germany or Europe to have that. Local Data-centers are bad, mmkay /s
- Sep 13
-
Hard position as not getting behind China is even more critical. Even if they agree to slow down I would not trust them. Advanced AI in Chinese hands is far more dangerous than advanced AI
- Sep 8
-
China-based AI companies are conducting industrial-scale distillation campaigns to extract proprietary capabilities from U.S. AI models. Our joint advisory with @NSAGov & @FBI details the threat and actions to help protect U.S. AI leadership. Read more 👉
-
This report sheds new light on the lengths Chinese AI companies will go to in order to fast-follow American AI, the sophistication of their techniques and the doggedness with which they carry out their activities, “likely with Chinese government awareness
-
China-based AI companies are illicitly distilling U.S. frontier AI capabilities. Read NSA's new report, co-sealed with @FBI, @CISAgov, highlighting AI knowledge distillation, TTPs used, and recommended mitigations: