DeepSeek's Sensitive Data Exposed on the Web

Cybersecurity firm Wiz has discovered a large amount of sensitive data from the popular AI company DeepSeek exposed on the Internet, raising concerns about security in the field of artificial intelligence.

Accordingly, the New York-based cybersecurity company (USA) discovered more than 1 million unsecured data lines of DeepSeek, including digital software keys and chat logs, which appear to record prompts sent from users to the company's free AI assistant.

DeepSeek caught inadvertently leaking data online.

Wiz said it immediately notified DeepSeek, and the Chinese company quickly fixed the issue in less than an hour. However, Wiz CTO Ami Luttwak cautioned that 'this was so easy to find that we believe we weren't the only ones who found it,' suggesting that other parties may have had access to the data before.

The incident comes as DeepSeek has emerged as a 'new star' in the AI ​​scene, competing directly with US tech giants like Nvidia and Microsoft. DeepSeek's ability to deliver powerful AI models at low cost has been impressive, but has also raised questions about the sustainability of the business model and profit margins of US AI companies.

The data breach has highlighted the importance of security in AI, especially as companies increasingly collect and process large amounts of user data. It also shows that even the most advanced tech companies can make security mistakes, and that these mistakes can have serious consequences.

DeepSeek has not yet commented on this incident.

Micah Soto

Update 10 February 2025