Anthropic says Chinese AI firms Moonshot and DeepSeek covertly routed user requests to its Claude models and then presented the resulting answers as if they came from their own systems. The company frames the allegation as part of a broader pattern of unfair competition in the AI sector.
Bloomberg reports that the dispute involves Moonshot and that Anthropic claims the firm routed thousands of user requests to Claude, producing responses that were delivered to users under Moonshot’s branding. The South China Morning Post similarly reports Anthropic’s accusations, adding that Anthropic describes the conduct as secret routing and passing off the outputs as proprietary.
The outlets both center on Anthropic’s allegation that these firms used Claude behind the scenes while maintaining the appearance of independent model performance. The reporting also points to a wider context of growing US scrutiny of China’s AI industry practices and related accusations between US and Chinese companies, though details about evidence and timelines are not fully established in the excerpts provided.
