On Thursday, Futurum Group CEO Daniel Newman accused Chinese AI labs of relying on U.S. frontier models to advance their own systems, after Anthropic disclosed large-scale efforts to extract capabilities from Claude.
Daniel Newman Slams China AI Narrative
Newman took to X, calling the alleged activity "freaking insane" and arguing that some Chinese AI models may not be as independently developed as widely portrayed.
"Now can we talk about why these Chinese "Open Weight" models are so good?" Newman wrote. He alleged that Chinese labs were "routing prompts through Claude" and using the resulting outputs to train their own models.
His comments followed a threat intelligence report from Anthropic detailing what it called "illicit distillation" involving several China-based AI companies, including Alibaba Group Holdings (NYSE:BABA), Moonshot and DeepSeek.
Anthropic Details Massive Claude Distillation
Distillation involves using responses from a more capable AI model to help train another system.
In a report released Thursday, Anthropic said the activity it identified went beyond ordinary model use and involved unauthorized, large-scale extraction of Claude’s capabilities.
Alibaba allegedly generated more than 151 million exchanges with Claude between May and July, making it the largest distillation campaign Anthropic said it had detected. The activity reportedly peaked at nearly 3 million exchanges per day across more than 3,500 fraudulent accounts.
Anthropic said Alibaba used Claude outputs to help train its Qwen models and also relied on the model for areas including reinforcement learning and model architecture research.
Moonshot, DeepSeek Also Accused
Anthropic said Moonshot AI, the company behind Kimi, routed some customer requests to Claude without users’ knowledge.
During one 10-day period, nearly 300,000 requests were allegedly relayed to Anthropic, while more than 23 million exchanges were attributed to Moonshot between May and July.
The company also accused DeepSeek of similar activity, saying it observed more than 12 million distillation attacks over 14 days in July.
Anthropic said some of the intercepted requests contained sensitive information and raised concerns about privacy and terms-of-service violations.
The report covered activity across multiple areas, including cyber operations, surveillance, scams, biological misuse, weapons development and AI model distillation.
Alibaba, Moonshot, DeepSeek and Xiaomi did not immediately respond to Benzinga’s request for comment.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Photo Courtesy: Algi Febri Sugita on Shutterstock.com
Login to comment