US accuses China AI developers DeepSeek and Alibaba of copying American AI
Key Points:
- The FBI, NSA, and CISA released a report accusing top Chinese AI companies of systematically mining advanced AI models from U.S. firms like Anthropic, OpenAI, Google, and xAI since 2024 to develop their own models through a practice called distillation.
- Distillation, while common in the AI industry, often violates AI companies’ terms of service, and U.S. officials allege that Chinese AI models heavily rely on this method to replicate American technologies.
- The report named six leading Chinese AI developers involved in this practice and suggested that the campaign likely had awareness from the Chinese government, though it did not directly implicate Chinese intelligence agencies.
- Chinese officials denied the allegations, emphasizing their AI development as a product of self-reliance and calling for increased U.S.-China cooperation rather than accusations.
- The report recommended U.S. companies improve user verification, share information on suspicious activities, and modify AI responses to suspected malicious queries, highlighting the broader concern over intellectual property theft and national security.