DeepSeek has taken the AI world by storm over the last few days, much to the chagrin of Silicon Valley, with Microsoft accusing the Chinese AI outfit of “improperly” obtaining OpenAI data to train its model.
The DeepSeek v3launch was so successful, in fact, that it led to tech shares across the US to tank in a major way, causing panic across the industry about what to do in regard to this new and open source competitor.
Microsoft has decided to go with the accusatory route, insisting that the only way the company could produce such impressive results at such a low cost is by stealing data from other established AI models like ChatGPT.
Suspicious Activity
The story was broken by Bloomberg News, which says that Microsoft and OpenAI are teaming up to launch the probe into whether DeepSeek improperly obtained data to train its AI model.
While an OpenAI spokesperson didn’t point the finger directly at DeepSeek, they did tell Reuters that US AI companies were constantly under siege from “adversaries and competitors” looking to take their IP. They also said that it is “critically important that we are working closely with the US government to best protect the most capable models.”
This just in! View
the top business tech deals for 2025 👨💻
The US government has also waded in. In an interview with Fox News, the White House’s AI and crypto lead, David Sacks, actually named DeepSeek. He stated: “There’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models.”
Gloves Are Coming Off
This probe marks how competitive the AI arms race is becoming. The US government is actively pushing AI innovation with Trump’s new Stargate AI program, with some biggest players in AI, including OpenAI, SoftBank, and Oracle, involved.
The program will see 20 data centers constructed across the US with the hope that it will create hundreds of thousands of jobs.
While AI innovation is being touted as a bid to drive economic growth, Trump’s frosty relationship with the Chinese government is also playing a part, creating a race to the top that his administration wants to win.
Not Exactly Angels
Before indignation takes hold, it’s worth pointing out that OpenAI itself has been the subject of accusations of improprieties including mining copyrighted materials.
Among the lawsuits was one in which both OpenAI and Meta were accused by comedian and author, Sarah Silverman, of summarizing her works without attribution. OpenAI argues that it needs copyrighted materials to train its LLMs and can’t afford to pay for them – an argument that does not go down well with content creators.
LLMs are voracious and need data. OpenAI — and possibly all of the AI innovators — have likely crossed lines in obtaining this data. But when it comes to a Chinese company potentially doing something improper, they are proving very quick to shout about it.
