Taiwan barred all of its government agencies and critical infrastructure service providers from using Chinese artificial-intelligence startup DeepSeek’s technology, citing security concerns.
“Its operation involves security concerns such as cross-border transmission and information leakage,” the self-governing island’s Ministry of Digital Affairs said in a statement on Friday.
DeepSeek in January released a new open-source AI platform that can mimic human reasoning, drawing awe and consternation internationally for its apparent competitiveness with world-leading AI bots.
DeepSeek’s model was developed and seemingly trained at a fraction of the cost of rivals’ platforms, suggesting that advanced AI apps could be created without enormous investments in hardware. The company’s free namesake app has also shot to popularity globally, topping mobile downloads in multiple countries.
But hurdles to DeepSeek’s growth are already emerging, as governments and businesses consider its potential cybersecurity implications and the possibility that data and other information collected by the Hangzhou-based company could be shared with China’s government.
Italy’s privacy regulator has blocked the app, the UK issued a warning to its citizens and businesses, while the US Pentagon restricted access. Hundreds of companies, including law firms, have also stopped their staff from accessing DeepSeek’s tools.
Taiwan’s digital ministry said that no official or confidential information should be used to query DeepSeek, given its Chinese origins. “It is a product that endangers national information security,” the ministry’s statement said.
{{/usCountry}}Taiwan’s digital ministry said that no official or confidential information should be used to query DeepSeek, given its Chinese origins. “It is a product that endangers national information security,” the ministry’s statement said.
{{/usCountry}}--With assistance from Debby Wu and Cindy Wang.
More stories like this are available on bloomberg.com
©2025 Bloomberg L.P.