Anthropic Accuses Alibaba of Stealing AI Capabilities Through 28.8 Million Fraudulent Exchanges in Largest Known Distillation Attack
Summary
Anthropic accuses Alibaba of orchestrating the largest known AI distillation attack in history, using 28.8 million exchanges across 25,000 fraudulent accounts to illicitly steal its AI capabilities, and is now urging the U.S. Senate to take coordinated action against the threat.
Key Points
- Anthropic sends a letter to the U.S. Senate accusing Alibaba of conducting the largest known distillation attack on its AI systems, using 28.8 million exchanges through roughly 25,000 fraudulent accounts between April 22 and June 5.
- Operators affiliated with Alibaba and its AI lab are accused of 'brazenly' and 'illicitly' attempting to extract Anthropic's AI capabilities through distillation, a method where a weaker model is trained using outputs from a stronger one.
- Anthropic is calling for coordinated government and industry action to combat illicit distillation, noting that Alibaba ignored prior warnings from the Trump administration, while Anthropic itself faces its own ongoing dispute with the administration over export controls on its latest AI models.