Microsoft Corp. and OpenAI are investigating a suspected unauthorized data extraction linked to DeepSeek, a Chinese artificial intelligence startup, raising concerns over the security of proprietary AI models.
According to sources familiar with the matter, Microsoft’s security researchers observed individuals they believe to be linked to DeepSeek extracting a large volume of data using OpenAI’s application programming interface (API).
The activity, detected in late 2024, prompted Microsoft to notify OpenAI about the potential breach.
If confirmed, the data extraction could violate OpenAI’s terms of service and suggest an attempt to bypass restrictions on how much data could be obtained from OpenAI’s proprietary models. Neither Microsoft nor OpenAI have officially commented on the investigation.
DeepSeek, which recently released its R1 AI model, has positioned itself as a major competitor to OpenAI, Google, and Meta Platforms, claiming that R1 rivals or outperforms leading U.S. AI models, particularly in mathematical tasks and general knowledge, while being developed at a fraction of the cost.
The emergence of DeepSeek has already had significant market implications, triggering a sell-off in AI-related stocks. Shares of Microsoft, Nvidia, Oracle, and Google-parent Alphabet fell sharply, collectively wiping out nearly $1 trillion in market value on Monday.
David Sacks, President Donald Trump’s AI czar, weighed in on the allegations, stating there is “substantial evidence” that DeepSeek leveraged OpenAI’s outputs to train its own AI models.
“There’s substantial evidence that what DeepSeek did here is they distilled knowledge out of OpenAI models, and I don’t think OpenAI is very happy about this,” Sacks said in an interview with Fox News, referencing a method called distillation, where one AI model is trained using the outputs of another.
In response to Sacks’ comments, OpenAI acknowledged the ongoing risks associated with its technology being targeted by foreign competitors.
“We know PRC-based companies — and others — are constantly trying to distill the models of leading US AI companies. As the leading builder of AI, we engage in countermeasures to protect our IP, including a careful process for which frontier capabilities to include in released models, and believe as we go forward that it is critically important that we are working closely with the US government to best protect the most capable models from efforts by adversaries and competitors to take US technology.”
DeepSeek, as well as High-Flyer, the hedge fund where the company was initially incubated, have not responded to requests for comment.
With the AI industry’s competitive landscape evolving rapidly, U.S. firms are increasingly ramping up security measures to safeguard their intellectual property from potential threats.
The latest allegations against DeepSeek highlight growing tensions between Western AI leaders and Chinese competitors, raising questions about the future of global AI development and regulation.