UK scrutinizes DeepSeek chatbot over national security fears as China accelerates AI race

British officials are examining the national security implications of DeepSeek, whose AI model has caused panic in Silicon Valley and sparked fears China has stolen a march in the global tech race.

The Chinese company released a model this month, called R1, to rival OpenAI’s ChatGPT, saying it was trained at a fraction of the cost. The free app has shot to the top of Apple’s App Store chart, ahead of ChatGPT and Google’s Gemini.

Speaking to POLITICO from Brussels on Wednesday, Britain’s Technology Secretary Peter Kyle said: “We scrutinize every innovation of the size and scale and impact of DeepSeek and we will make sure that it goes through the right system.

“We have a very mature intelligence and security apparatus in the United Kingdom.

"It’s a very regular occurrence that new technologies, new products will emerge onto the global economy and I just want to reassure people in Britain … the system that we have will look at this as it does at every other innovation and make sure that safety is there from the onset."

Kyle did not detail the exact nature of the probe, but Britain’s National Cyber Security Center — a division of the GCHQ surveillance agency — is known to scan for future technological risks.

The development has already sparked enquiries from data regulators in Italy, Ireland and Australia and drawn the attention of the U.S. national security apparatus.

Italy’s data regulator said Tuesday it had sent DeepSeek requests to disclose what personal data it collects, where it comes from, how it’s used, what legal basis it has to process that data under the European Union’s General Data Protection Regulation and whether that data is stored on servers in China.

There are growing concerns amongst German data protection authorities over the app too, with Zeit reporting that authorities are investigating potential regulatory steps — starting with a formal inquiry into the company’s data processing practices.

Research from NewsGuard, a U.S.-based company which provides publishers with accuracy ratings, placed DeepSeek one from bottom in its veracity tests of similar chatbots. It also found the chatbot acted as a “mouthpiece for China” in some of its responses. 

But Kyle argued there were some positives to DeepSeek’s rise, saying it showed the importance of innovation — and linked it to his Cabinet colleague Rachel Reeves’ flurry of growth-focused announcements on Wednesday.

“That’s why in the U.K. … we’ve launched the AI action plan, why we are building up our digital capacity, why we are… announcing a link between Cambridge and Oxford universities. We’re going to create one of the biggest clusters in the world of innovation,” he said.

Chris Lunday contributed to this report.