Microsoft and OpenAI are investigating whether data from OpenAI’s technology was improperly acquired by a group linked to Chinese AI startup DeepSeek, Bloomberg News reported on Tuesday.
According to the report, Microsoft’s security researchers noticed individuals they believed to be associated with DeepSeek exfiltrating large amounts of data in the fall using OpenAI’s application programming interface (API). OpenAI’s API is the primary method for developers and business clients to access its services.
Microsoft, which is the largest investor in OpenAI, alerted the company about the suspicious activity, the report said.
DeepSeek, a low-cost Chinese AI startup challenging U.S. competitors, caused a tech stock selloff on Monday after its free AI assistant surpassed OpenAI’s ChatGPT on Apple’s App Store in the U.S.
David Sacks, the White House’s AI and crypto czar, told Fox News earlier on Tuesday that it was “possible” DeepSeek had stolen intellectual property from the U.S., adding, “There’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models.”
When asked for comment on the Bloomberg report, an OpenAI spokesperson affirmed Sacks’ statement, noting that China-based companies and others are continually trying to replicate the models of leading U.S. AI companies. However, the spokesperson did not specifically mention DeepSeek or any other company.
“We implement counter-measures to protect our intellectual property, including a thorough process to decide which advanced capabilities to include in released models,” the spokesperson said. “As we move forward, it’s crucial to collaborate with the U.S. government to protect the most advanced models from efforts by adversaries and competitors to steal U.S. technology.”
Microsoft declined to comment, and DeepSeek could not be reached for immediate comment.
