DeepSeek’s arrival has shaken the foundations of the current artificial intelligence industry. However, its data handling policies have also raised red flags among experts and US officials alike. Now, it appears that “hundreds” of companies and government organizations are blocking DeepSeek over concerns about data security.
“Hundreds” of companies blocking DeepSeek AI on their systems, cybersecurity firms say
Cybersecurity firms such as Armis and Netskope have revealed that many of their clients have requested DeepSeek blocks on their systems. “The biggest concern is the AI model’s potential data leakage to the Chinese government,” Armis’ CTO Nadir Izrael told Bloomberg. “You don’t know where your information goes.”
On the other hand, Ray Canzanese, director of Netskope’s threat labs, revealed that 52% of his clients have requested to set access restrictions to DeepSeek AI from their systems. The numbers at Armis are even higher, reaching 70% of clients. The vast majority of companies making such requests are related in one way or another to government entities.
San Francisco law firm Fox Rothschild also took similar measures, Bloomberg Law reports. In the US, both the Pentagon and the Navy have begun blocking access to the platform.
Government organizations around the world are also taking action
Internationally, Italy ordered the blocking of DeepSeek “as a matter of urgency and with immediate effect.” Ireland’s Data Protection Commission requested information from the Chinese company on the user data protection.
In the UK, the Information Commissioner’s Office only issued a “warning.” The agency recalled that AI platforms present in the country must be transparent about their user data handling policies. If not, they could take action at any time to correct it. In the US, officials and experts have warned about the risks posed by storing data on Chinese servers. Chinese laws force companies to comply with government requests for private user data.
Recently, cybersecurity startup Wiz found a dataset from DeepSeek publicly available without any protection. It contained information such as chat histories, backend details and technical log data from the platform. PromptFoo, another similar startup, found that DeepSeek is vulnerable to jailbreaks. It also proved that the service avoids responding to most prompts related to “sensitive topics” for China.
Potential growth opportunity for cybersecurity software developers
That said, some see DeepSeek’s emergence as a potential opportunity. Growing data privacy concerns could increase interest in acquiring more powerful cybersecurity tools. Bloomberg Intelligence reports that companies such as CrowdStrike Holdings Inc., Palo Alto Networks Inc. and SentinelOne could benefit from this.
