Skip to content
Tech News
← Back to articles

Chinese AI labs secretly used millions of Claude exchanges to train their models, Anthropic says

2026-09-11 | original
read original get The Coming Wave by Mustafa Suleyman (book) → more articles
Why This Matters

Anthropic is publicly accusing major Chinese AI labs — Alibaba, Moonshot and DeepSeek — of using fraudulent accounts to harvest Claude outputs at massive scale to train competing models, a practice it calls 'illicit distillation.' If accurate, it suggests frontier model capabilities can be cheaply cloned through API access, undermining the moat of expensive pretraining. It also raises privacy questions, since Anthropic says user and corporate data was silently routed to Claude without disclosure.

Key Takeaways
Worth a Look

The Coming Wave by Mustafa Suleyman (book) — If stories about rival AI labs quietly copying each other's models make you want the bigger picture, this book digs into how fast frontier AI is spreading and why containing it is so hard. It's a readable primer on the competitive and geopolitical pressures behind headlines like Anthropic's distillation report.

See The Coming Wave by Mustafa Suleyman (book) on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.

Anthropic said it detected and disrupted unauthorized large-scale efforts by China-based AI labs including Alibaba , Moonshot and DeepSeek to train their models using Claude.

The U.S. AI company said in a threat intelligence report released Thursday that the firms were involved in what it described as "illicit distillation," a process in which outputs from a more capable AI model are used to train another model and replicate some of its capabilities without authorization.

"Some of these exchanges included sensitive information, including from individual users, major multinational companies, and state-affiliated actors ... These practices are likely inconsistent with privacy laws and the labs' own terms of service," according to the report.

Anthropic said operators affiliated with Alibaba used Claude outputs to help train its Qwen models, while Moonshot routed some Kimi user requests to Claude and used some of the resulting exchanges to train its own models.

The Alibaba operation was the largest distillation campaign Anthropic said it has measured, involving more than 151 million exchanges with Claude between May and July.

The activity peaked at nearly 3 million exchanges per day from more than 3,500 fraudulent accounts, according to the report. The company said Alibaba also used Claude for broader AI research, including reinforcement learning and model architecture.

Moonshot and DeepSeek

Anthropic also detailed activity involving Moonshot AI, the Beijing-based company behind the Kimi family of AI models.

Moonshot silently forwarded some customer requests intended for Kimi to Claude and then displayed Claude's responses to users, who thought they were using a Kimi model, according to the report.

In one 10-day period, Moonshot relayed nearly 300,000 customer requests to Anthropic, the vast majority of which were routed to Claude Opus models. The requests were routed through a network of 5,380 accounts that Anthropic described as fraudulent, most of which appeared to be located in Singapore and Japan.

... continue reading

Explore topics: anthropic claude alibaba moonshot ai deepseek
Related:
Y Combinator’s Garry Tan says 'do nothing' about distillation as AI giants accuse China of copying their tech
Show HN: Algo-Trading-Skills - 501 agent skills for trading infrastructure
Anthropic Says It Blocked Possible Efforts to Build Biological Weapons
Anthropic caught scientists using Claude to further biological weapon research
Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek

© 2026 GoKawiil

About | Editorial Policy | Contact