Chinese police arrest man who allegedly used ChatGPT to spread fake news

Chinese police arrested a man they allege used ChatGPT to create a fake news article about a train crash, in what appears to be the first case of enforcement action being taken in China under an unprecedented law related to artificial intelligence.

The case highlights Chinese authorities’ push to regulate and control uses of AI as the technology gets more advanced.

Police in Gansu province in northwest China detained a man, surnamed Hong, who they said allegedly fabricated a news story regarding a train crash that caused nine deaths.

The authorities found that more than 20 accounts had posted this article on a blogging platform owned by Chinese search giant Baidu and they’d garnered more than 15,000 views.

Hong allegedly used ChatGPT to create slightly different versions of the fake news article to pass duplication checks on the Baidu-owned platform.

ChatGPT, created by U.S. firm OpenAI, is an example of an chatbot based on generative AI technology, which allows software to generate responses based on user prompts and questions. For example, users can ask ChatGPT to come up with a story based on specific instructions.

The Gansu police authorities arrested Hong under the first-of-its kind law governing “deep synthesis technologies” which China introduced this year. Deep synthesis technologies refer to AI being used to generate text, images, video or other media. The law states that deep synthesis services cannot be used to disseminate fake news.

China drafted the law as ChatGPT was taking off and going viral, as authorities looked to get ahead of the technology. China’s internet is heavily censored and controlled. Beijing has sought to introduce laws governing new technologies which could present concerns to the central government.

ChatGPT is blocked in China but can be accessed with the use of a virtual private network — a software that can help bypass the country’s internet restrictions.

Chinese technology giants are currently trialing their own rivals to ChatGPT. But these are not widely-available chatbots like ChatGPT. Instead, Chinese firms have been more cautious in their approach and targeted specific uses in part, analysts told CNBC, not to spook regulators and the government.

For example, Alibaba’s Tongyi Qianwen AI product will eventually be rolled out on its DingTalk workplace communication software, and Tmall Genie, a provider of smart home appliances.