---
title: "Anthropic Disrupts Russian Cyber Espionage and Chinese AI Model Theft"
publication: "MediaBias News"
url: "https://mediabias.news/business/anthropic-disrupts-russian-cyber-espionage-and-chinese-ai-model-theft"
api: "https://mediabias.news/api/v1/stories/anthropic-disrupts-russian-cyber-espionage-and-chinese-ai-model-theft"
markdown: "https://mediabias.news/business/anthropic-disrupts-russian-cyber-espionage-and-chinese-ai-model-theft.md"
audio: "https://mediabias.news/api/audio/anthropic-disrupts-russian-cyber-espionage-and-chinese-ai-model-theft"
category: "Business & Economy"
area: "China"
published: "2026-09-11T13:46:09.941Z"
source_reported: "2026-09-11T12:22:55.179Z"
updated: "2026-09-11T13:46:09.941Z"
trust_score: 60
critic_score: 65
hype_score: 45
assessment_type: "coverage-and-source-reporting-analysis"
fact_check_status: "not-performed"
coverage_measured: "2026-09-11T14:45:11.885035+00:00"
---

# Anthropic Disrupts Russian Cyber Espionage and Chinese AI Model Theft

*AI firm reports thwarting state-linked hacking and efforts by Chinese companies to replicate its Claude models.*

**Scores for the source reporting** (0-100, assessing the original journalism this article was written from, not this write-up): 

- Trust 60 of 100, higher is better. Central claims attributed to Anthropic's report, with two full reports and multiple independent outlets corroborating.
- Craft 65 of 100, higher is better. Reporting clearly separates Anthropic's claims from analysis, and details specific companies and figures.
- Hype 45 of 100, lower is better. Headlines and digests use strong language like 'disrupts' and 'accuses' but the body is factual.

Scored by MediaBias News; method at https://mediabias.news/methodology.

**The short version**

- Anthropic reported disrupting a suspected Russian cyber-espionage campaign linked to the SVR, which targeted Ukrainian organisations.
- Seven China-based AI labs, including Alibaba, Moonshot, DeepSeek, and Xiaomi, are accused of attempting to extract capabilities from Claude.
- Alibaba allegedly used over 151 million exchanges with Claude between May and July 2026, employing more than 3,500 fraudulent accounts.
- Moonshot and DeepSeek are accused of routing live customer conversations through Claude to use the responses as training data.

Anthropic has disrupted several attempts to misuse its Claude AI models. The company reported thwarting a suspected Russia-linked cyber-espionage campaign and efforts by Chinese AI firms to extract and replicate Claude's capabilities. These actions were detailed in Anthropic's latest Threat Intelligence report.

The Russia-linked campaign was associated with Midnight Blizzard, a threat actor previously linked by the US government to the SVR, Russia's Foreign Intelligence Service. This group allegedly targeted Ukrainian government, military, and diplomatic organisations. Anthropic stated that AI was used across several stages of these operations, including phishing and malware evasion.

## Disagreement on specific numbers of exchanges

Reports differ on the exact number of exchanges attributed to specific Chinese AI labs. Anthropic stated that Alibaba's alleged campaign involved over 151 million exchanges with Claude between May and July 2026, using more than 3,500 fraudulent accounts, as reported by EM360 and Quartz. However, the-decoder.de reported that Alibaba's Qwen team alone had more than 151 million exchanges. Business Insider reported that China's star AI labs routed user requests to Claude at least 35 million times in the summer, while hngn.com stated that seven Chinese AI labs ran nearly 200 million exchanges to extract Claude's capabilities. Cryptopolitan reported that accounts linked to Alibaba ran more than 151 million exchanges with Claude to train its Qwen models, calling it the largest distillation campaign documented.

## What the coverage left out

None of the left-rated, centre-rated, or right-rated reports below mention Anthropic's claim that it disrupted activity associated with affiliates of the ShinyHunters cybercrime collective. Additionally, none of the reports mention Anthropic's claim that it disrupted activity involving operators using the model to develop software associated with conventional weapons, including firearms, missiles, armed drones, and other munitions, or additional activity involving weapons design, intelligence gathering, and procurement linked to programmes in China, Russia, and Yemen.

## Who covered it

Shares of the 17 covering outlets with a published leaning rating:

- Left: 18% (3)
- Centre: 58% (10)
- Right: 24% (4)

25 of 42 covering outlets carried no usable leaning rating and were excluded from those percentages. Coverage measured 2026-09-11T14:45:11.885035+00:00.

**Coverage watch:** developing. Checked 7 times; 0 checks found a material change. Most recently checked 2026-09-11T15:30:13.277Z.

## How the sides framed it

### Left

- The left-rated reports focused on the alleged actions of Chinese AI labs. CNBC reported that Chinese AI labs secretly used millions of Claude exchanges to train their models, with Alibaba and Moonshot among those involved in unauthorized "distillation" aimed at extracting capabilities from Claude. Business Insider highlighted that China's star AI labs routed user requests to Claude at least 35 million times in the summer, naming Alibaba, Moonshot, Zhipu, DeepSeek, and Xiaomi as attempting to use Claude to train their models in various ways.

### Centre

- Centre-rated reports detailed Anthropic's accusations against Chinese AI labs for illicit distillation. Quartz reported that Alibaba's campaign alone involved over 151 million exchanges with Claude between May and July. South China Morning Post stated that Anthropic claims Moonshot and DeepSeek secretly diverted user requests to Claude. Reuters reported that Anthropic disrupted bioweapons research efforts, Russian hacking, and Chinese Claude misuse. TechCrunch detailed distillation campaigns from Alibaba, Moonshot AI, and DeepSeek, noting that competition in the space has intensified. Bloomberg reported that Moonshot secretly routed user requests through Claude. Memeburn noted Anthropic's distillation battle has spread into an illicit market of fake accounts, proxies, and stolen credentials.

### Right

- Right-rated reports highlighted Anthropic's disruption of Russian and Chinese AI campaigns. Times of India reported that Anthropic disrupts Russian, Chinese AI campaigns targeting its Claude models, revealing disruption of cyber espionage campaigns linked to Russia and Chinese firms exploiting its Claude AI models. CNBC-TV18 stated that China's Alibaba ran the largest AI 'brain-theft' operation ever recorded, extracting over 151 million Claude exchanges to help train Qwen models. Yonhap News Agency reported on Anthropic's 'Distillation' report, stating that Moonshots and others replicated capabilities through 'transfer stations' using fake accounts.

## Factuality profile of the covering outlets

These are published factuality ratings of the outlets, not a verdict on whether this story or its claims are true.

- mixed: 5
- high: 8
- unknown: 26
- veryHigh: 3

## Verification scope

This page compares coverage and scores the source reporting. It is not a ClaimReview verdict on whether the underlying event or claim is true.

## Original reporting this was written from

- [The Decoder](https://the-decoder.com/how-hackers-used-claude-for-missiles-drone-swarms-and-surveillance-while-chinese-labs-mined-it-for-training-data) — How hackers used Claude for missiles, drone swarms, and surveillance, while Chinese labs mined it for training data
- [Portal Mix Vale](https://mixvale.com.br/2026/09/11/ai-distillation-anthropic-details-how-alibaba-deepseek-and-moonshot-use-its-models) — AI Distillation: Anthropic Details How Alibaba, DeepSeek, and Moonshot Use Its Models
- [shiftdelete.net](https://shiftdelete.net/anthropic-acikladi-cinli-yapay-zeka-sirketlerine-guvenmeyin) — Anthropic Announces! Don't Trust Chinese AI Companies!
- [Quartz](https://qz.com/anthropic-chinese-ai-labs-distillation-alibaba-deepseek-moonshot-091126) — Anthropic accuses Chinese AI labs of illicit distillation attacks
- [L'Humanité](https://humanite.fr/monde/anthropic/intelligence-artificielle-anthropic-accuse-ses-concurrents-chinois-de-voler-ses-modeles) — Artificial Intelligence: Anthropic Accuses Chinese Competitors of Stealing His Models
- [the-decoder.de](https://the-decoder.de/raketen-drohnenschwaerme-ueberwachung-ersatz-fuer-china-modelle-anthropic-zeigt-wofuer-claude-missbraucht-wird) — Rockets, Drone Swarms, Surveillance, Replacement for China Models: Anthropic Shows What Claude Is Being Abused For
- [economiematin.fr](https://economiematin.fr/anthropic-denonce-pillage-claude-chine-echec-autoregulation) — Anthropic Denounces the Plundering of Claude by China and the Failure of Self-Regulation
- [EM360](https://em360tech.com/tech-articles/anthropic-disrupts-russian-chinese-ai-campaigns-targeting-claude) — Anthropic Disrupts Russian, Chinese Campaigns Targeting Claude AI
- [Times of India](https://enterpriseai.economictimes.indiatimes.com/news/industry/anthropic-disrupts-russian-chinese-ai-campaigns-targeting-its-claude-models/134057817) — Anthropic disrupts Russian, Chinese AI campaigns targeting its Claude models
- [Unknown](https://hani.co.kr/arti/international/china/1277437.html) — Antropic “Alibaba Approached Claude 151 Million Times”… China’s ‘Malicious Distillation’ Case Revealed
- [Aroged](https://aroged.com/2026/09/11/anthropic-found-deepseek-and-moonshot-redirecting-customer-requests-to-claude-instead-of-their-own-models) — Anthropic found DeepSeek and Moonshot redirecting customer requests to Claude instead of their own models
- [Noticias ao Minuto](https://noticiasaominuto.com/tech/3049714/anthropic-acusa-rivais-chinesas-de-copiarem-capacidades-do-claude) — Anthropic Accuses Chinese Rivals of Copying Claude's Capabilities
- [technology.org](https://technology.org/2026/09/11/anthropic-cuts-off-seven-chinese-ai-labs-and-a-russian-spy-crew) — Anthropic Cuts Off Seven Chinese AI Labs and a Russian Spy Crew - Technology Org
- [t3nMagazin](https://t3n.de/news/anthropic-china-claude-firmen-ki-1762722) — Anthropic Reports: Kimi Used 5,380 Fake Accounts to Tap Claude
- [Business Insider](https://businessinsider.com/china-ai-labs-millions-distillation-attacks-anthropic-claude-2026-9) — China's star AI labs routed user requests to Claude at least 35 million times in the summer: Anthropic
- [CNBC-TV18](https://cnbctv18.com/technology/chinas-alibaba-ran-the-largest-ai-brain-theft-operation-ever-recorded-anthropic-report-19989124.htm) — China's Alibaba ran the largest AI ‘brain-theft’ operation ever recorded: Anthropic report
- [Crypto Briefing](https://cryptobriefing.com/anthropic-report-warns-ai-distillation-boosts-reasoning-poses-security-risks) — Anthropic report warns AI distillation boosts reasoning, poses security risks
- [Cryptocurrency News | Cryptocurrency Prices | Market Cap](https://cryptocurrency.com.tr/anthropic-report-warns-ai-distillation-boosts-reasoning-poses-security-risks-373755) — Anthropic report warns AI distillation boosts reasoning, poses security risks
- [it-daily.net](https://it-daily.net/shortnews/anthropic-china-taeuschung) — Anthropic Accuses Chinese AI Providers of Deception
- [INTELLINEWS](https://intelli.news/2026/09/11/kunstliche-intelligenz-anthropic-wirft-chinesischen-ki-rivalen-schummelei-vor) — Künstliche Intelligentz: Anthropic Wirft Chinesichen KI-Rivalen Schummelei Vor
- [Los Angeles Weekly Times](https://losangelesweeklytimes.com/chinese-ai-labs-secretly-used-millions-of-claude-exchanges-to-train-their-models) — Chinese AI labs secretly used millions of Claude exchanges to train their models
- [Cryptopolitan](https://cryptopolitan.com/anthropic-alibaba-claude-distillation) — Anthropic ties 151 million Claude queries to alleged Alibaba distillation
- [Unknown](https://chosun.com/economy/tech_it/2026/09/11/GQ2TIMDFMQZDOMRZMM3TAZRVMI) — Antropic: “China’s Moonshot AI Bypassed Claude Access with Thousands of Fake Accounts… Unauthorized Distillation Detected”
- [unsafe.sh](https://unsafe.sh/go-441614.html) — How Chinese AI Companies Are Attempting to Clone American AI Models
- [hngn.com](https://hngn.com/articles/273187/20260910/anthropic-accuses-7-chinese-ai-labs-copying-claude-scale.htm) — Anthropic Names Alibaba, DeepSeek in AI Distillation Report
- [CNBC](https://cnbc.com/2026/09/11/chinese-ai-labs-moonshot-deepseek-alibaba-anthropic.html) — Chinese AI labs secretly used millions of Claude exchanges to train their models, Anthropic says
- [PYMNTS.com](https://pymnts.com/news/artificial-intelligence/2026/anthropic-reports-chinese-labs-clone-claude-capabilities-on-industrial-scale) — Anthropic Reports Chinese Labs Clone Claude Capabilities on Industrial Scale
- [ddaily.co.kr](https://ddaily.co.kr/page/view/2026091108113318367) — Did Claude Copy It Secretly?… Antropic Raises Allegations of Distillation by 7 Chinese AI Sites
- [cybernoz.com](https://cybernoz.com/moonshot-deepseek-secretly-routed-user-requests-to-claude-anthropic-claims) — Moonshot, DeepSeek secretly routed user requests to Claude, Anthropic claims
- [South China Morning Post](https://scmp.com/news/us/diplomacy/article/3367112/moonshot-deepseek-secretly-routed-user-requests-claude-anthropic-claims) — Anthropic claims Moonshot, DeepSeek secretly diverted user requests to Claude
- [연합뉴스-Yonhap News Agency](https://yna.co.kr/view/AKR20260911014800009) — Antropic 'Distillation' Report... "Moonshots and Others Replicated Through 'Transfer Stations'"
- [BizToc](https://biztoc.com/x/0834ef481f18c612) — Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek
- [Reuters](https://reuters.com/legal/litigation/anthropic-disrupts-russian-chinese-ai-campaigns-targeting-its-claude-models-2026-09-10) — Anthropic disrupts bioweapons research efforts, Russian hacking, Chinese Claude misuse
- [TechCrunch](https://techcrunch.com/2026/09/10/anthropic-details-distillation-campaigns-from-alibaba-moonshot-ai-and-deepseek) — Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek
- [Wall Street Journal](https://wsj.com/tech/ai/chinese-ai-giants-accused-of-sending-millions-of-user-queries-to-u-s-models-768c9d26) — How Chinese AI Firms Tried to Clone U.S. AI Models
- [anthropic.com](https://anthropic.com/threat-intelligence-report-september-2026) — Countering misuse of AI: September 2026 / Anthropic
- [cisa.gov](https://cisa.gov/news-events/cybersecurity-advisories/aa26-251a) — China-Based Artificial Intelligence Companies Conducting Industrial-Scale Distillation Campaigns Against U.S. AI Companies
- [Perfil](https://perfil.com/noticias/bloomberg/bc-anthropic-acusa-a-moonshot-de-desviar-consultas-a-claude-para-entrenar-su-propia-ia.phtml) — Anthropic Accuses Moonshot of Diverting Consultations From Claude to Train His Own AI
- [1prime.ru](https://1prime.ru/20260910/kompanii-873206239.html) — Chinese Companies Secretly Redirected User Requests to Claude
- [DiarioBitcoin](https://diariobitcoin.com/china/anthropic-acusa-a-moonshot-ai-de-desviar-miles-de-solicitudes-hacia-claude) — Anthropic Accuses Moonshot AI of Diverting Thousands of Requests to Claude
- [Bloomberg](https://bloomberg.com/news/articles/2026-09-10/moonshot-secretly-routed-user-requests-through-claude-anthropic-says) — Moonshot Secretly Routed User Requests Through Claude, Anthropic Says
- [Memeburn](https://memeburn.com/anthropics-distillation-battle-moves-to-the-dark-web-as-china-fears-grow) — Anthropic’s Distillation Battle Moves To The Dark Web As China Fears Grow

## Questions

**What is AI model distillation?**

AI model distillation is a process where the outputs from a more capable AI model are used to train another, often smaller, model. This allows the smaller model to mimic some of the capabilities of the larger one without needing to develop them independently.

**Which Chinese AI labs are accused of model distillation?**

Anthropic has accused seven China-based AI labs of attempting to extract capabilities from Claude. Among those named are Alibaba, Moonshot, DeepSeek, and Xiaomi. These companies allegedly used Claude's outputs to train and improve their own AI models.

**What was the scale of Alibaba's alleged activity?**

Anthropic claims that Alibaba's alleged campaign involved over 151 million exchanges with Claude between May and July 2026. This activity reportedly used more than 3,500 fraudulent accounts and peaked at nearly three million exchanges per day.

**What other misuse of AI models did Anthropic report?**

Anthropic also reported disrupting a suspected Russia-linked cyber-espionage campaign targeting Ukrainian organisations. The company stated that AI was used across various stages of these operations, including phishing and malware evasion.

---

MediaBias News — https://mediabias.news. Reproduced from https://mediabias.news/business/anthropic-disrupts-russian-cyber-espionage-and-chinese-ai-model-theft. Please cite as: MediaBias News, "Anthropic Disrupts Russian Cyber Espionage and Chinese AI Model Theft", https://mediabias.news/business/anthropic-disrupts-russian-cyber-espionage-and-chinese-ai-model-theft