In context: AI startup DeepSeek stunned the world with the release of its R1 AI model, which can mimic human reasoning at a level comparable to the best OpenAI models to date. While the company has received widespread acclaim for its achievement, it is now reportedly under investigation by Microsoft and OpenAI for allegedly accessing OpenAI's data output illegally to train its AI model.
According to unnamed sources cited by Bloomberg, the probe started last fall after Microsoft's security researchers discovered that a group linked to DeepSeek had accessed a large amount of data through OpenAI's API. While developers can legally pay for a license to use this data in their applications, Microsoft suspects that DeepSeek's actions violated OpenAI's terms of service.
Following the discovery, Microsoft informed OpenAI that it had detected an attempt by a DeepSeek-affiliated group to bypass restrictions on how much data a single party can access. The two companies then launched a joint investigation into the incident, which is now being treated as a potential data breach.
An OpenAI spokesperson declined to confirm the alleged data theft but told Reuters that the company employs cutting-edge "countermeasures" to protect its intellectual property. They added that OpenAI is working closely with the U.S. government to prevent its AI models from falling into the hands of foreign adversaries and competitors.
However, the White House's AI and crypto czar, David Sacks, was less restrained in his response to the report. In an interview with Fox News, he claimed there was "substantial evidence" that DeepSeek had accessed OpenAI's data in an unethical and possibly illegal manner. Microsoft refused to comment, while DeepSeek could not be reached for its reaction on the controversy.
With the launch of its all-conquering AI model, DeepSeek appears poised to challenge OpenAI, Google, and Meta in the field of AI research. However, if the allegations turns out to be accurate, it could spell trouble for the Chinese AI firm, which saw its new app overtake ChatGPT to become the top free app on both the App Store and Play Store in the U.S. this week.
DeepSeek also received unexpected praise from President Trump this week when he described the release of the R1 LLM as a "positive" development and an "asset" for the global tech sector. He noted that if DeepSeek delivers on its promise of accelerating AI training at a lower cost, "that's good (and) I view that as a positive."
Microsoft is investigating if DeepSeek illegally accessed OpenAI data to train its AI model
