David Sacks, Trump’s AI and crypto “czar,” said in an interview on Fox on Tuesday that there’s “substantial evidence” that Chinese AI company DeepSeek “distilled” knowledge from OpenAI’s AI models, a process that Sacks compared to theft.

Sacks, who didn’t cite the source of this “evidence,” suggested that DeepSeek used responses from OpenAI models to train its own. “I don’t think OpenAI is very happy about this,” Sacks said.

DeepSeek, which went viral this week for its openly available, highly performant models and free chat apps, is being looked at with increasing scrutiny by U.S. government officials. According to Reuters, the National Security Council — the U.S. body that advises the president on foreign policy and national security — is reviewing the implications of DeepSeek’s apps, which topped the Apple App Store this past weekend.

The U.S. Navy has also reportedly banned the use of DeepSeek’s AI, citing “potential security and ethical concerns.”