Data Privacy Concern a Red Flag as DeepSeek AI Stores User Data in China

By Tenzin Chokyi

DHARAMSALA 30 Jan: Chinese artificial intelligence(AI) startup DeepSeek, which has triggered a storm in the US over its advanced AI capabilities, has proven to be a state-controlled AI model designed to align with and reinforce the Chinese government’s official narratives, particularly on politically sensitive issues such as Tibet, Taiwan, Tiananmen square and its border dispute with India, among others.

The Chatbot’s refusal to address certain topics has raised concerns about the influence of China in shaping narratives online, especially within the Tibetan exile community. This becomes even more pressing as Tibet remains directly under China’s colonial rule.

Lobsang Gyatso Sither, Director of Technology at the Tibet Action Institute (TAI), expressed serious concerns about the spread of pro-Chinese narratives regarding Tibet and East Turkestan on Deepseek. 

Sither says in a video post on TAI’s Instagram handle that this issue is particularly serious, explaining that AI operates differently from traditional search engines like Google or social media platforms. While those platforms offer a variety of perspectives for users to analyse, AI typically provides direct answers, limiting the opportunity for users to explore multiple viewpoints.

He has told RFA that the ability of AI to shape public perception in alignment with the Chinese government’s ideological framework presents both ethical and security challenges.

Geshe Lobsang Monlam, founder and CEO of the Monlam Tibetan IT Research Centre, also says in the report that the worldwide free AI service by China is essentially to establish and spread pro-Chinese narratives.

DeepSeek, like many apps developed in China, stores user data on servers located in China. As such, it may be subject to Chinese law, potentially allowing access by the Chinese government under certain circumstances. While DeepSeek gathers and stores user data similar to other AI companies like ChatGPT and Claude, it is important to note that the specifics of its data retention practices are less clear. 

The app collects technical information such as device model, operating system, IP address, and usage patterns.

Earlier in 2023, China introduced the interim measure for the management of Generative Artificial Intelligence Services(AI Measures), which requires AI-generated content to align with China’s core socialist values and standards. The measure also grants the government significant oversight over data, user activity and algorithms, strengthening state surveillance and control over digital spaces.

Sparking concerns about state surveillance,  Australia’s Science Minister, Ed Husic, raised privacy issues regarding DeepSeek, as reported by the BBC

He has said that he would exercise caution, emphasising that such matters need to be carefully considered. 

The US Navy has reportedly banned its members from using DeepSeek citing “potential security and ethical concerns,” according to CNBC.

The app is reportedly not available in Italy due to concerns over data protection.  

Leave a Reply

Your email address will not be published. Required fields are marked *