Chinese AI models are priced low, and British businesses and consumers are willing to accept them. However, in the view of The Guardian in the UK, this has become a “temptation” that the UK needs to be vigilant against. While admitting that Chinese AI models are entering the market at prices far lower than those of American competitors, he also calls on the UK to refuse to use them. He also claims that Western AI systems at least have “good intentions” in terms of security, whereas Chinese AI does not have such intentions.
On October 17th, the British newspaper The Guardian published a commentary by economic columnist Phillip Inman, titled "Why Britain Should Reject Cheap and Terrible Chinese AI."
Yingman acknowledged that Chinese companies such as DeepSeek and Xiaomi continuously launch new open-source AI models, which users can deploy and modify on their own, without the need to pay expensive ongoing software licensing fees.
He also mentioned that British consumers prefer cheap Chinese products, and Chinese electric vehicles are an example of this. However, he claimed that the Chinese government promotes car exports through subsidies, thereby weakening foreign car industries. The so-called threat posed by cheap Chinese AI is even greater.
Yingman believes that Chinese companies are launching a large number of AI models at a fraction of the cost compared to their American competitors. For users, they have the option to choose affordable and customizable Chinese open-source models, rather than purchasing expensive American software that comes with many unnecessary features.
But he then changed the subject, stating that there is an important difference between Chinese and American AI products: in daily use, Western AI makes certain efforts to ensure security and does so with “good intentions,” whereas Chinese AI does not have these conditions.
The article quotes Ross Burley, director of the British non-profit organization 'Center for Information Resilience', stating that even running the DeepSeek model on local servers does not mean that political censorship tendencies can be eliminated or security vulnerabilities can be fixed.
The article also cites a 2025 evaluation of DeepSeek by the National Institute of Standards and Technology (NIST), criticizing the models for political censorship issues.
However, Inman's affirmation of the “good intentions” of Western AI contrasts sharply with the safety issues that have been exposed in recent years regarding American AI products.
A report released by the British Artificial Intelligence Safety Institute (AISI) in December 2025 indicates that researchers have found "generalized ways to bypass security protection" in all advanced AI systems tested, allowing models to output content that was previously restricted. The report also notes that the security capabilities of different models vary, and some new systems require more complex attacks to be breached.
In August this year, OpenAI in the United States also publicly admitted that during an internal cybersecurity assessment conducted in July, its AI models bypassed network isolation measures and accessed OpenAI's internal research infrastructure, as well as some systems of the US AI development platform Hugging Face, without authorization.
These models were in a testing environment where safety restrictions have been relaxed. They cannot be directly compared to products that are available to the public and have the same issues. However, this incident still exposed the security risks of advanced AI systems in the United States.
Just two days before the publication of this article in The Guardian, Reuters reported on October 15, citing Axios, that the U.S. Federal Trade Commission had issued civil subpoenas to OpenAI and Anthropic, requiring the two companies to submit documents related to security incidents, protective measures, and consumer rights protection.
At the end of the article, Inman still insists that Britain should resist the attraction of cheap Chinese AI, strengthen regulation of American AI, and support domestic and EU suppliers. He warns that reliance on Chinese technology, including in the automotive and AI sectors, will hollow out British industry and threaten data security.
