OpenAI and Anthropic warn Washington about Chinese distillation of U.S. AI models
The Neuron ● Covered by 3 sources
Anthropic and OpenAI accused Chinese companies including Alibaba and DeepSeek of using a technique called "distillation" to copy their AI models without permission and produce cheaper alternatives. Anthropic reported that Alibaba generated 28.8 million outputs from Claude using nearly 25,000 fraudulent accounts between April and June 2024. The theft allows Chinese AI firms to narrow the technology gap from an estimated 18-plus months behind to 6-9 months, enabling them to undercut US companies on price and potentially erode American AI leadership.
Why it matters
OpenAI and Anthropic warned U.S. officials that Chinese companies are using distillation at enormous scale to clone advanced American models, potentially converting billions of dollars in research into shortcuts for rival labs.
