Anthropic accuses Alibaba of largest distillation attack on Claude
It sounds like a movie plot, but it’s real: Anthropic publicly accused Alibaba of orchestrating the largest distillation attack ever seen against Claude.
What happened?
Between April and June 2026, a scheme involving over 25,000 fraudulent accounts generated 28.8 million interactions with Claude. The goal? Steal the model’s knowledge through distillation — basically, using Claude’s responses to train a competing model.
Distillation, explained
Distillation is when you ask millions of questions to an AI model and use the answers to train another one. Like an athlete copying a rival’s moves in slow motion.
Anthropic claims to have clear evidence linking the scheme to Alibaba, which hasn’t commented yet.
This case could become a major legal precedent for AI intellectual property disputes.