Microsoft and OpenAI are reportedly investigating whether data from the artificial intelligence firm’s tech stack was improperly obtained by a group linked to Chinese AI startup DeepSeek.
Microsoft security researchers detected suspicious large-scale data extraction through OpenAI’s application programming interface (API) in late 2024, Bloomberg reported on Jan. 29, citing people familiar with the matter.
Microsoft, OpenAI’s largest investor, notified the firm of the activity, which could violate the ChatGPT maker’s terms of service or could mean the group removed restrictions on how much data they could gather.
On Jan. 20, the China-based DeepSeek released its latest AI model, R-1, which reportedly rivals market leader ChatGPT’s performance with a much lower build cost. The announcement caused a tech and AI stock slump that wiped billions from the US market.
The White House crypto and AI czar, David Sacks, told Fox News on Jan. 28 that there is evidence DeepSeek used OpenAI’s model outputs to train their own AI through a process called distillation — where an AI model uses the outputs of another for training purposes to develop similar capabilities.
“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.
AI Czar David Sacks says American companies will learn efficiency techniques from China's DeepSeek AI model, but big AI data centers are still needed and scaling the biggest data centers is still an advantage pic.twitter.com/SuSlhGKO3J
— Tsarathustra (@tsarnick) January 28, 2025
OpenAI acknowledged general concerns about Chinese companies attempting to distill US AI models but didn’t specifically address DeepSeek.
“We know PRC-based companies — and others — are constantly trying to distill the models of leading US AI companies,” an OpenAI spokesperson told Bloomberg, referring to the People’s Republic of China.
The investigation centers on possible violations of OpenAI’s terms of service or potential circumvention of data access restrictions and highlights growing tensions around AI technology competition between the US and China.
Related: Crypto mining stocks loss extend, tech stocks steady after DeepSeek scare
Meanwhile, CNBC reported on Jan. 28 that the US Navy banned its members from using DeepSeek out of fear the Chinese government could exploit sensitive data.
The Navy sent a warning email to personnel on Jan. 24 stating that the AI model was not to be used “in any capacity” due to “potential security and ethical concerns associated with the model’s origin and usage.”
Magazine: You should ‘go and build’ your own AI agent: Jesse Pollak, X Hall of Flame