A comment piece in The Guardian argues that Britain should turn down cheap artificial intelligence models built in China, such as DeepSeek, because of risks to data security and the political censorship built into them. The column calls on the UK to work on taming US AI and to back domestic and EU suppliers instead.
What happened
The Guardian has published a viewpoint column on its business pages saying that AI developed in China and offered to Western businesses and governments almost free of charge gives good reason for concern. The piece notes that Chinese developers such as DeepSeek and Xiaomi release new versions of their “open-weight” models every month. These models let users take the core software and run it without paying an expensive ongoing licence.
The column compares this with the debate over Chinese electric cars, which it describes as cheaper than other EVs sold in the UK. It points out that Beijing subsidises those cars to boost exports, and argues that cheap AI takes the same export threat further. In the writer’s view, AI has given the Chinese state much greater surveillance powers at home, and lets Beijing build dependence abroad by making other countries rely on Chinese technology.
The piece describes the choice facing an AI user as one between a cheap Chinese open system that can be adapted, and expensive US software from companies such as OpenAI and Anthropic. It says both carry serious risks at the extreme. In day-to-day use, though, it argues that only the US option makes at least some effort on security and benign intent. It also mentions that Meta is developing open models, and that some tech analysts say European and Indian versions are in development.
The Guardian quotes Ross Burley, director of the Centre for Information Resilience (CIR), an independent non-profit group whose work includes exposing human rights abuses and countering disinformation. Burley says running DeepSeek in isolation deals with only part of the risk. “Running a model on your own servers doesn’t automatically remove any political censorship. Or indeed repair weaknesses in any security defences,” he says. He adds: “We should be deeply concerned about embedding Beijing’s political restrictions into systems that businesses and public services rely on for information and decisions.”
The column cites an evaluation report by the US National Institute of Standards and Technology (Nist). The report found political censorship in downloaded DeepSeek models running on Nist’s own infrastructure. According to The Guardian, this means hosting those models in the UK would not remove the censorship by itself.
Why it matters if you live in United Kingdom
If you use DeepSeek’s online chatbot, the piece describes what its policy allows. According to The Guardian, the policy says the service collects prompts and uploaded material, processes and stores personal data in China, and permits disclosure to Chinese authorities. Michael Wooldridge, a professor of AI at the University of Oxford, told the newspaper it is not unreasonable to assume that data entered into the chatbot could be shared with the Chinese state.
For residents, this is a practical point. Personal details, work documents or questions typed into the chatbot may not stay in the UK. If you have moved here from abroad and use these tools for job applications, visa paperwork, tenancy documents or anything else sensitive, it is worth reading the privacy terms of any AI service before you upload files.
If you work for a UK business or public body, the column raises a second issue. Organisations sometimes download open-weight models and run them on their own servers to save money and keep data in-house. Burley and the Nist findings suggest this does not remove the political censorship built into those models. Where such a model is used to supply information or support decisions, those restrictions could affect what staff and the public are told.
For consumers, the piece links AI with the wider debate about cheap Chinese products in Britain, including electric cars. Its argument is that relying on Chinese technology, from cars to AI, could hollow out British industry and weaken the security of people’s data. This is the writer’s opinion rather than government policy, but it reflects a debate that could shape what products and services are available to UK buyers.
What to watch next
The column ends with a direct policy call: the UK should reject cheap Chinese AI, seek to better tame US AI, and support domestic and EU suppliers. Whether ministers or regulators act along these lines is the main thing to watch. Any move could affect which AI tools public services and businesses are allowed or encouraged to use.
It is also worth watching the European and Indian models the piece says are in development, and Meta’s open models. If they become competitive, users who want low-cost, adaptable AI may get alternatives that do not raise the concerns described here.
Finally, Chinese developers continue to release models every month, according to The Guardian. Each new release is likely to bring fresh questions about data handling and censorship. Further independent testing, such as the evaluation Nist carried out, will be important for anyone deciding which tools to trust.
Source: The Guardian
Sources
- The Guardian: Why Britain should reject the lure of cheap – and scary – Chinese AI, accessed 18 October 2026
The UK, for internationals
One email a week: the week's news for internationals in the UK, new practical guides and what changed in the cities. No spam, unsubscribe any time.
