Microsoft and OpenAI are investigating whether data output from ChatGPT was accessed without authorization by a group associated with Chinese AI startup DeepSeek. According to Bloomberg News, Microsoft's security researchers observed individuals linked to DeepSeek extracting a significant amount of data through OpenAI's API last fall.
OpenAI's API is the principal method for developers and business customers to access its services. Being the primary investor in OpenAI, Microsoft promptly alerted the company about the suspicious activity.
DeepSeek, a budget-friendly alternative to U.S. competitors, gained attention in the tech industry as its free AI assistant surpassed OpenAI's ChatGPT on Apple's App Store in the U.S. This development raised concerns, with White House AI and crypto czar David Sacks hinting in a Fox News interview that DeepSeek may have illicitly acquired intellectual property from the U.S.
Sacks mentioned, "There's substantial evidence that DeepSeek extracted knowledge from OpenAI's models."
OpenAI, in response to the Bloomberg report, emphasized the ongoing attempts by Chinese companies and others to replicate leading U.S. AI models. While not directly naming DeepSeek or any specific company, the spokesperson stressed the importance of collaboration with the U.S. government to safeguard advanced models from exploitation.
Microsoft chose not to provide a comment, and DeepSeek was unavailable for immediate comment.