Summary

Microsoft (MSFT, Financials) and OpenAI are investigating whether individuals linked to the Chinese artificial intelligence startup DeepSeek improperly accessed OpenAI's proprietary data, according to Bloomberg sources.

Last October, Microsoft's security researchers saw unusual behavior by people allegedly connected to DeepSeek exfiltrating significant volumes of data over OpenAI's application programming interface, or API. Terms of service from OpenAI demand developers follow data consumption limits and pay for API access. The episode begs questions about possible policy violations by OpenAI or an effort to circumvent use restrictions.

Recently, DeepSeek introduced R1, an artificial intelligence model that has apparently surpassed American competitors at much reduced prices. The news coincided with a Monday selloff in the $1 trillion technology industry.

AI czar in former President Donald Trump's government David Sacks told Fox News there is "substantial evidence" DeepSeek extracted information from OpenAI's models using a machine learning method known as distillation. Distillation is the process of spreading information from bigger artificial intelligence systems to smaller ones, hence perhaps allowing copycat models to develop.

Sacks said that U.S. AI companies are anticipated to adopt further protections against model distillation not too far off. OpenAI said it keeps working with the U.S. government to safeguard intellectual property and admitted that China-based businesses have before tried to condense American artificial intelligence algorithms.