U.S. accuses China’s DeepSeek of AI intellectual property theft

U.S. accuses China’s DeepSeek of AI intellectual property theft
The DeepSeek logo is seen in this illustration taken on 29 January, 2025.
Reuters

The United States has issued an international warning accusing Chinese firms, including AI start-up DeepSeek, of allegedly stealing intellectual property from American artificial intelligence labs. 

A diplomatic cable dated Friday (24 April) and seen by Reuters was sent to U.S. diplomatic and consular posts worldwide, instructing officials to raise concerns with foreign counterparts about what it described as the “extraction and distillation of U.S. AI models” by adversaries.  

The document said a separate diplomatic démarche had also been sent to Beijing.  

“Distillation” refers to the process of training smaller AI models using outputs from larger, more advanced systems, often to reduce development costs.  

The State Department did not immediately respond to a request for comment from Reuters. 

CEO of OpenAI Sam Altman speaks during the 2026 Infrastructure Summit of government officials, corporate executives, and labor leaders, in Washington, D.C., U.S., 11 March 2026.
Reuters

OpenAI, which developed ChatGPT, has previously warned U.S. lawmakers that DeepSeek was attempting to replicate leading American AI models for its own systems.  

China rejected the allegations. The Chinese embassy in Washington said the claims were “groundless” and accused the U.S. of attempting to hinder China’s technological progress.  

DeepSeek on Friday launched a preview of a new model, V4, adapted for chips produced by Chinese technology company Huawei.  

The company did not immediately respond to Reuters’ request for comment.  

Previously, DeepSeek said its V3 model relied on publicly available data collected through web crawling and denied intentionally using synthetic data generated by OpenAI.  

Several Western governments, along with some in Asia, have restricted the use of DeepSeek by officials and institutions over data privacy concerns.  

The State Department cable said its aim was to warn partners about the risks of using AI systems derived from what it described as unauthorised distillation of U.S. proprietary models.

It added that such systems may appear competitive on certain benchmarks at lower cost but do not match the full capabilities of the original models, and may remove safeguards designed to ensure neutrality and reliability.  

Tags