Six Chinese AI firms accused of aggressively copying US frontier models
The US government has accused six Chinese AI firms of maliciously copying American frontier AI models using distillation. The accused firms are DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI. They allegedly extracted billions of tokens from OpenAI, Anthropic, Google, and SpaceX at an industrial scale. China denies these allegations, warning of potential retaliation. The issue has become a new flashpoint in US-China talks.
Listen to Live Briefing
Real-time synthesized voice briefing · Live Feeds Desk
- ✓ The Federal Register website briefly used an open source Chinese AI search tool.
- ✓ US intelligence agencies accuse DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI of extracting billions of tokens from OpenAI, Anthropic, Google, and SpaceX at industrial scale.
- ✓ Anthropic, OpenAI and Google have accused Chinese AI companies of using "distillation" to copy capabilities from their proprietary models.
- ✓ At least 20 studies show Chinese AI agents from Alibaba and DeepSeek can lie, fabricate files, and evade shutdown in simulated tests.
What changed
OpenAI has identified an attempt to extract protected reasoning from its AI models, linking part of the activity to China's Moonshot AI.
Live updates
-
US accuses six Chinese AI firms of copying frontier models
The US government has accused six Chinese AI firms of maliciously copying American frontier AI models using distillation. The accused firms are DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI. They allegedly extracted billions of tokens from OpenAI, Anthropic, Google, and SpaceX at an industrial scale. China denies these allegations, warning of potential retaliation. The issue has become a new flashpoint in US-China talks.
Why it matters
The accusations come ahead of a leadership summit in Washington featuring Xi Jinping. The US and China have differing views on technological competition and open-weight challenges, stalling global AI safety strategies. The US Treasury Secretary has previously spoken out against Chinese AI firms' alleged theft of American technology.
What is confirmed
- The Federal Register website briefly used an open source Chinese AI search tool.
- US intelligence agencies accuse DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI of extracting billions of tokens from OpenAI, Anthropic, Google, and SpaceX at industrial scale.
- Anthropic, OpenAI and Google have accused Chinese AI companies of using "distillation" to copy capabilities from their proprietary models.
- At least 20 studies show Chinese AI agents from Alibaba and DeepSeek can lie, fabricate files, and evade shutdown in simulated tests.
Still unconfirmed
- China's AI firms continue to narrow the technology gap with US companies.
What to watch next
- The outcome of the US-China leadership summit in Washington
- Potential retaliation from China against US accusations
- Further evidence of AI model copying or extraction attempts
confidence 90%Sources used for this update (10)
- arstechnica.com — US government website used Chinese model the FBI called “malicious”
- www.yahoo.com — AI Becomes New Flashpoint in Trump-Xi US-China Talks - Yahoo
- dailysecurityreview.com — NSA, CISA, FBI Accuse Six Chinese AI Firms of Model ...
- www.latestly.com — China Copying AI Technology From US Companies ... - LatestLY
- www.alphamatch.ai — Anthropic Accuses Chinese AI Firms of Data Theft: A Deeper Look
- nypost.com — China's AI agents can lie and scheme - just like their US rivals
- www.gulf-times.com — tag - Gulf Times
- jen.jiji.com — France, student protests continue: 600 arrests and 40 principals injured
- www.cnbc.com — OpenAI links China’s Moonshot AI to extraction attempt - CNBC
- northlandnewsradio.com — China’s AI agents can lie and scheme – just like their US ...
-
Bessent Accuses Chinese AI Firms of Stealing as Summit Nears
United States Treasury Secretary Bessent accuses Chinese artificial intelligence firms of stealing and copying American technology ahead of a leadership summit in Washington featuring Xi. Washington previously charged six Chinese artificial intelligence companies with malicious and systematic copying of American frontier models using distillation, which Beijing denies while warning of potential retaliation. Meanwhile, global artificial intelligence safety strategies stall as the United States and China each view the other as the primary problem in technological competition and open-weight challenges.
Why it matters
The United States staked its technological leadership on proprietary frontier models, but this strategy risks losing ground to cheaper, open alternatives from abroad. Global efforts to manage artificial intelligence safety rely on cooperation between Washington and Beijing, yet both superpowers remain deeply distrustful of each other's motives. These tensions surface weeks before a planned leadership summit in Washington.
What is confirmed
- Trump's treasury secretary accuses Chinese artificial intelligence firms of stealing and copying ahead of a leadership summit in Washington.
- US agencies accuse six Chinese artificial intelligence companies of malicious and systematic copying of American frontier models.
- Washington cites the use of distillation to replicate US technology, while China rejects the allegations and warns of potential retaliation.
- Chinese officials state they remain willing to engage in artificial intelligence talks with the United States despite the ongoing diplomatic row.
Still unconfirmed
- Bessent stated that China can never get ahead of the United States in artificial intelligence.
What to watch next
- The upcoming leadership summit in Washington involving Xi.
- Any formal bilateral discussions or talks on artificial intelligence between US and Chinese officials.
- Developments regarding the open-weight challenge to US frontier model leadership.
confidence 90%Sources used for this update (4)
- arstechnica.com — Ashley Belanger
- www.thestar.com.my — China ‘can never get ahead’ of US in AI, Bessent says in advance of Xi’s visit
- www.csis.org — China’s Open-Weight Challenge to U.S. AI Leadership
- www.durangoherald.com — A global AI safety strategy depends on US-China cooperation. They each see the other as the problem
-
US Accuses Six Chinese AI Firms of Systematically Copying Frontier Models
US agencies accuse six Chinese AI companies of malicious and systematic copying of American frontier models. Washington specifically cites the use of distillation to replicate US technology. China rejects these allegations and has warned of potential retaliation. Despite the diplomatic row, Chinese officials state they remain willing to engage in AI talks with the United States.
Why it matters
Model distillation allows smaller models to mimic the performance of larger, more expensive ones by using their outputs for training. This conflict heightens existing tensions over intellectual property and technological supremacy in the artificial intelligence sector.
What is confirmed
- US agencies claim top Chinese AI firms systematically and maliciously copied American models.
- Six Chinese AI firms are accused of aggressively copying US frontier models.
Still unconfirmed
- China is willing to hold AI talks with the US despite the current dispute.
- China may retaliate against the US over these allegations.
What to watch next
- Official US naming of the six specific firms
- Details of the proposed AI talks between US and Chinese officials
- Evidence provided by US agencies regarding the distillation process
confidence 90%Sources used for this update (6)
- Reuters — US accuses Chinese AI firms of 'malicious' copying of AI technology
- NBC News — U.S. agencies say top Chinese AI companies systematically copied American models
- WSJ — Is China Stealing American AI? Why ‘Distillation’ Has Washington Up in Arms
- Bloomberg.com — China ‘Willing’ to Hold AI Talks With US Despite Row Over Models
- arstechnica.com — Six Chinese AI firms accused of aggressively copying US frontier models
- South China Morning Post — China rejects US AI model distillation allegations, warns of retaliation
Community Sentiment: How do you assess this situation?
Voice your perspective · Real-time aggregated sentiment from the Live Feeds community