What's happened
Anthropic reports several Chinese firms have engaged in distillation attacks, using outputs from advanced models to train less capable ones. Alibaba, Moonshot, DeepSeek and others are implicated, with millions of exchanges rerouted to Claude and used to train rival systems. The firm reports dozens of millions of affected interactions and warns of regulatory and security risks.
What's behind the headline?
Short Analysis
- Distillation attacks are escalating as firms push cheaper, faster AI development.
- The practice erodes user privacy and could violate terms of service, regulators will scrutinize governance.
- If validated, this could shift the competitive landscape toward firms that can secure data and prevent data leakage.
What this means for readers
- Expect more robust monitoring of AI data flows and identity verification for high-risk regions.
- Businesses should reassess data-sharing agreements with outsourcing partners.
Forecast
- Regulators may tighten compliance requirements; insurers may reassess AI liability risk.
How we got here
The report aggregates activity from December 2025 to August 2026, detailing how Alibaba, Moonshot, DeepSeek, Xiaomi and others have rerouted user prompts to Claude and used those responses to train alternative models. Anthropic backs its findings with threat intelligence and accounts of fraudulent activity.
Our analysis
- New York Post: Anthropic alleges distillation attacks by Alibaba, Moonshot, DeepSeek and others. The report highlights 151 million exchanges and 3,500 fraudulent accounts. - Business Insider UK: Details on 190 million distillation attacks and government data exposure. - Al Jazeera: Reports on weaponization concerns and state-aligned actors.
Go deeper
- What safeguards are in place to prevent distillation attacks?
- Which companies are most at risk for regulatory action?
- How should users protect sensitive personal data when interacting with AI services?
More on these topics
-
Anthropic - Artificial intelligence company
Anthropic PBC is a U.S.-based artificial intelligence startup public-benefit company, founded in 2021. It researches and develops AI to "study their safety properties at the technological frontier" and use this research to deploy safe, reliable models for
-
People's Republic of China - Country in East Asia
China, officially the People's Republic of China, is a country in East Asia. It is the world's most populous country, with a population of around 1.4 billion in 2019.
-
Japan - Country in East Asia
Japan is an island country of East Asia in the northwest Pacific Ocean. It borders the Sea of Japan to the west and extends from the Sea of Okhotsk in the north to the East China Sea and Taiwan in the south.
-
Singapore - Country in Asia
Singapore, officially the Republic of Singapore, is a sovereign island city-state in maritime Southeast Asia. It lies about one degree of latitude north of the equator, off the southern tip of the Malay Peninsula, bordering the Straits of Malacca to the w
-
United States - Country in North America
The United States of America, commonly known as the United States or America, is a country mostly located in central North America, between Canada and Mexico.
-
Iran (Islamic Republic of Iran) - Country in the Middle East
Iran, also called Persia, and officially the Islamic Republic of Iran, is a country in Western Asia. It is bordered to the northwest by Armenia and Azerbaijan, to the north by the Caspian Sea, to the northeast by Turkmenistan, to the east by Afghanistan a
-
Xiaomi - Consumer electronics company
Xiaomi Corporation is a Chinese electronics company founded in April 2010 and headquartered in Beijing. Xiaomi makes and invests in smartphones, mobile apps, laptops, bags, earphones, shoes, fitness bands, and many other products.