Chinese AI labs secretly used millions of Claude exchanges to train their models, Anthropic says
Source: CNBC Tech (opens in a new tab)
Intel Summary
Anthropic stated it detected unauthorized efforts by China-based artificial intelligence labs, including Alibaba and Moonshot AI, to harvest millions of Claude conversational exchanges to train and refine their own competing AI models.
Why It Matters
The incident underscores the growing industry challenge of preventing unauthorized model distillation and scraping by competitors, heightening intellectual property concerns and competitive tensions between leading Western AI developers and Chinese labs.
Part of an ongoing development
Independent reportingAnthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek
TechCrunch reports that Anthropic has published a report alleging persistent model distillation campaigns conducted by China-based AI companies, specifically citing Alibaba, Moonshot AI, and DeepSeek. Claims are as reported; this summary makes no determination about accuracy or significance.
- Confidence
- High confidence
- Corroboration
- Corroborated
More coverage of this development
- Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeekTechCrunchIndependent reporting
Organizations & Entities
Related Intelligence
- ReportAlso involving Moonshot AI
Hackers Used Claude and Grok to Fully Automate Attacks on Banks
According to an investigation by cybersecurity startup Straiker reported by The Information, a hacking group named CyberCartel deployed a network of AI agents powered by models from Anthropic, SpaceX, and Moonshot AI to automate cyberattacks against more than 400 Latin American financial targets. The research represents one of the first documented cases of threat actors using multi-model agentic AI systems to successfully automate intrusions into financial institutions.
The Information - ReportAlso involving Claude
AI Guardrails Aren’t Stopping Criminal Hackers
WIRED reports that threat actors globally are deploying agentic AI systems to execute cyberattacks. Malicious actors are actively bypassing safety guardrails and manipulating commercial models, specifically citing Anthropic's Claude, to facilitate and conduct criminal operations.
WIRED - ReportAlso involving Claude
How Bank Hackers Got Claude to Do Their Dirty Work
The Information reports that bank hackers have used Anthropic's Claude models to automate cyberattacks against financial targets. While Anthropic previously positioned its Mythos model as specialized for cybersecurity, the reporting indicates threat actors are successfully weaponizing standard Claude models without requiring Mythos-tier capabilities.
The Information - ReportAlso involving Claude
Anthropic's Claude now writes code and proves it safe automatically
The Decoder reports that Anthropic utilized its Claude model to automatically generate mathematical proofs verifying code correctness across 123 open-source libraries. The initiative generated approximately 180 million lines of proof code and identified previously undetected bugs that standard testing had missed, advancing Anthropic's objective to make AI-driven formal verification routine.
The Decoder