Governments across Asia are increasingly restricting access to DeepSeek, a Chinese AI chatbot, over concerns about national security. South Korea is the latest country to block DeepSeek on government-related networks, prohibiting its use of work computers within various ministries, including defence, trade, finance, and foreign affairs, as well as within its police force. Earlier in the week, Taiwan and Australia imposed complete bans on the chatbot across all government devices, citing risks related to security and privacy. The Australian government specifically stated that its decision was based on advice from national security agencies, warning that data uploaded to the app may not remain private and could expose users to potential malware threats.
India has also taken a cautious approach by prohibiting government employees from using AI tools such as DeepSeek and ChatGPT on official devices. However, it has not yet implemented an outright ban. The Indian government has justified these restrictions by emphasizing the risks posed to sensitive government documents and data confidentiality. Beyond Asia, concerns over DeepSeek’s data practices have been raised by European nations, including France and Italy. Italy, one of the first countries to take action, has launched an investigation into DeepSeek’s AI model, known as R1. These developments reflect growing skeptism over Chinese AI applications and their potential implications for data security.
The Chinese government has firmly denied allegations that DeepSeek is collecting unauthorised data. It has stated that neither the government nor any regulatory body compels companies or individuals to collect or store data unlawfully. Despite these reassurances, analysts warn that the crackdown on DeepSeek may be just the beginning of broader restrictions on generative AI tools, regardless of their country of origin. Some argue that concerns should not be limited to Chinese AI applications but should also extend to AI models developed by Western companies, as private corporations may have their interests that could pose risks to government security.
The escalating restrictions on AI tools signal a potential fragmentation of the global AI landscape, mirroring the divide seen in internet governance. Similar to how China developed its internet ecosystem, separate from the Western internet, the AI industry may be heading toward a future where Chinese and Western AI models operate in largely separate spheres with minimal competition or cross-border integration. Experts also highlight the increasing role of AI in cyber threats, with generative AI being used to tailor malware and enhance cyberattacks. As concerns over AI security grow, governments worldwide will likely adopt stricter regulations to safeguard sensitive information and mitigate emerging cyber risks.
This summary is generated from the YouTube video from CNA dated 7 Feb 2025.
So what has your government done? I like to hear it in the comments.
Update on 9 Feb 2025
Based on the article: “DeepSeek iOS app sends data unencrypted to ByteDance-controlled servers“, NowSecure, a mobile security company, raised the alarm that the DeepSeek app on iOS “sends sensitive data over unencrypted channels, making the data readable to anyone who can monitor the traffic. More sophisticated attackers could also tamper with the data in transit.” This is in contrast to what Apple had their developers do to enforce the encryption of data when it is sent over the wire using App Transport Security (ATS). What is more appalling is that the data is being sent to servers controlled by ByteDance, the Chinese company that owns TikTok. For more details, please reference the article.
No responses yet