A report has emerged that corporations and government agencies around the world are blocking access for their members to an AI chatbot developed by the Chinese artificial intelligence (AI) startup DeepSeek.
According to Bloomberg News on the 30th, Nadir Izraeel, chief technology officer (CTO) of the cybersecurity firm Armis, noted that hundreds of corporations, particularly those linked to the government, are blocking their employees’ access to DeepSeek due to concerns over potential data leaks to the Chinese government and vulnerabilities in personal information protection.
Izraeel noted that about 70% of their client corporations have requested to block access to DeepSeek.
The network security firm Netskope, which corporations use to block access to specific websites for their employees, also reported that 52% of its client corporations have completely blocked access to DeepSeek.
Izraeel said, “The biggest concern is the possibility of data leaking to the Chinese government,” adding, “We cannot tell where our information goes.”
DeepSeek states in its privacy terms that it collects and stores data on servers in China, and informs that disputes related to this will be subject to Chinese government law.
Bloomberg explained that DeepSeek can collect users’ keyboard patterns, text, audio, files, feedback, chat records, and other content input by users for training its AI model and may share that information with law enforcement agencies and public organizations at its discretion.
Cybersecurity experts investigating DeepSeek’s security have reported discovering publicly accessible databases of DeepSeek. The cybersecurity startup Wiz stated that this database contains some DeepSeek chat records, backend details, and log data.
The U.S. Congress has also restricted DeepSeek functionality on its assets and instructed employees not to install DeepSeek on public phones and computers.
Congress stated in a notice, “Threat actors are already exploiting DeepSeek as a means to distribute malicious software and infect devices.”
Bloomberg reported that prior to blocking access to DeepSeek on the 28th, the U.S. Department of Defense confirmed that some employees had used DeepSeek.