Israeli cybersecurity firm Wiz has uncovered a major security lapse in the infrastructure of Chinese AI startup DeepSeek. According to a blog post published Wednesday, Wiz detected that DeepSeek had inadvertently exposed over a million lines of sensitive data to the open internet, including digital software keys and chat logs capturing user prompts to its AI assistant.
Wiz's Chief Technology Officer, Ami Luttwak, revealed that DeepSeek acted swiftly to secure the exposed data after being notified. "They took it down in less than an hour," he said. However, he warned that the data was so easily accessible that others might have already found it.
DeepSeek, which has rapidly gained popularity with its AI assistant, did not immediately respond to requests for comment. The company’s success has captivated China while raising concerns in the U.S. about its potential to rival OpenAI at a fraction of the cost. The competitive pricing and performance of DeepSeek’s technology have fueled debates over the sustainability of American AI firms like Microsoft (NASDAQ:MSFT) and Nvidia (NASDAQ:NVDA).
By Monday, DeepSeek had surpassed OpenAI’s ChatGPT in downloads from Apple’s (NASDAQ:AAPL) App Store, triggering a selloff in global tech stocks. The incident highlights the security risks associated with AI startups as they scale rapidly, raising questions about data protection and regulatory oversight.
This discovery underscores the importance of robust cybersecurity measures for AI firms handling vast amounts of user data. As competition in the artificial intelligence sector intensifies, securing sensitive information will be critical for maintaining trust and avoiding potential breaches.