WASHINGTON, D.C.: The U.S. State Department has directed diplomats worldwide to raise concerns with foreign governments over what it describes as attempts by Chinese companies to extract and replicate American artificial intelligence models, according to a diplomatic cable seen by Reuters.

The cable, sent to U.S. embassies and consulates, instructs officials to engage counterparts on "concerns over adversaries' extraction and distillation of U.S. A.I. models," highlighting fears that proprietary systems developed by American firms are being copied at lower cost.

A separate message has been sent to Beijing to formally raise the issue with Chinese authorities, the document said.

The warning focuses on companies including DeepSeek, which has drawn global attention for its low-cost AI models, as well as firms such as Moonshot AI and MiniMax.

Distillation, the process at the center of the dispute, involves training smaller AI systems using outputs from larger, more advanced models to reduce development costs.

The move follows similar accusations from the White House earlier this week. The State Department did not immediately respond to requests for comment.

U.S.-based OpenAI has previously warned lawmakers that DeepSeek was attempting to replicate models developed by leading American firms. Reuters reported in February that such concerns had been raised during congressional discussions.

China has rejected the allegations. The Chinese Embassy in Washington said claims that its companies are stealing U.S. AI intellectual property are "groundless" and amount to "deliberate attacks on China's development and progress in the AI industry."

DeepSeek did not comment on the latest accusations. The company has previously said its models rely on publicly available data collected through web crawling and do not intentionally use synthetic data generated by rivals.

Despite growing scrutiny, DeepSeek's models remain widely used on global platforms that host open-source AI systems. However, several Western and some Asian governments have restricted the use of these tools among officials due to data privacy concerns.

The diplomatic cable said it aims to "warn of the risks of utilizing AI models distilled from U.S. proprietary AI models" and to prepare the ground for further U.S. government action.

It also argued that such models can appear competitive on certain benchmarks while lacking the full capabilities of the original systems. The document added that these efforts may remove built-in safeguards, including mechanisms designed to ensure neutrality and reliability.

The dispute comes amid intensifying competition in artificial intelligence, with major technology companies expected to invest more than US $600 billion in AI infrastructure this year.

The issue could further strain U.S.-China relations ahead of President Donald Trump's planned visit to China to meet President Xi Jinping.

Tensions between the two countries over technology and trade had eased following a détente last year, but the latest accusations risk reopening a key front in their ongoing rivalry over AI leadership.