Truace
Subscribe

Why Britain should reject the lure of cheap – and scary – Chinese AI

There is much to fear from artificial intelligence developed in China and handed to businesses and governments in the west almost free of charge. Every month Chinese AI developers such as DeepSeek and Xiaomi dispatch the latest versions of their “open-weight” models that allow users to take the core software and use it without paying an expensive ongoing licence. In Britain, we love cheap Chinese kit, usually setting aside any concerns about how it was made and who by. The row over whether to allow Chinese elect…

By the Truvace desk2 min read1 sourceJournalismTRV-2026-1704· cite & verify
Why Britain should reject the lure of cheap – and scary – Chinese AI

In brief

A Guardian opinion column argues that Chinese developers such as DeepSeek and Xiaomi are releasing open-weight AI models every month at very low cost. Users can adopt the core software without paying an ongoing licence. The piece compares this to cheap Chinese electric cars and frames both as an export threat.

The column matters because it presents AI model choice as a security and policy question for the UK, not only a price question. Its case rests on opinion and on a few cited sources, including a NIST evaluation, a DeepSeek privacy policy and academic comment. It also links AI to broader geopolitical claims that the text does not substantiate in detail.

Main points

  1. Chinese developers release open-weight models by the dozen each month at a fraction of the cost of US equivalents, according to the column.
  2. A US National Institute of Standards and Technology evaluation reportedly found political censorship in downloaded DeepSeek models running on its own infrastructure.
  3. DeepSeek's online chatbot policy says prompts and uploaded material are collected and personal data is stored in China, with disclosure to Chinese authorities permitted.
  4. The author recommends that the UK reject cheap Chinese AI, better tame US AI and support domestic and EU suppliers.

The problem

Embedding cheap Chinese open-weight models in UK businesses and public services risks carrying over political censorship and exposing user data to Chinese authorities.

The rundown

The column contrasts a cheap Chinese open system that users can tweak for their own ends with expensive US software that comes with features they may not need. It says both types carry potential dangers, but that only the US option comes with at least some effort to provide security and benign intent.

Ross Burley of the Centre for Information Resilience says that running a model on your own servers does not automatically remove political censorship or repair security weaknesses. He warns against embedding Beijing's political restrictions into systems that businesses and public services rely on.

Michael Wooldridge, an Oxford AI professor, says it is not unreasonable to assume that data entered into the DeepSeek chatbot could be shared with the Chinese state. The author concludes that relying on Chinese technology will hollow out industry and undermine data security.

Sources

  1. JournalismThe Guardian2026-10-17

The debate