Police in China have detained a man they say used ChatGPT to create fake news and spread it online, in what state media has called the country’s first criminal case related to the AI chatbot.
According to a statement from police in the northwest province of Gansu, the suspect allegedly used ChatGPT to generate a bogus report about a train crash, which he then posted online for profit. The article received about 15,000 views, the police said in Sunday’s statement.
ChatGPT, developed by Microsoft (MSFT)-backed OpenAI, is banned in China, though internet users can use virtual private networks (VPN) to access it.
Train crashes have been a sensitive issue in China since 2011, when authorities faced pressure to explain why state media had failed to provide timely updates on a bullet train collision in the city of Wenzhou that resulted in 40 deaths.
Gansu authorities said the suspect, surnamed Hong, was questioned in the city of Dongguan in southern Guangdong province on May 5.
“Hong used modern technology to fabricate false information, spreading it on the internet, which was widely disseminated,” the Gansu police said in the statement.
“His behavior amounted to picking quarrels and provoking trouble,” they added, explaining the offense that Hong was accused of committing.
Police said the arrest was the first in Gansu since China’s Cyberspace Administration enacted new regulations in January to rein in the use of deep fakes. State broadcaster CGTN says it was the country’s first arrest of a person accused of using ChatGPT to fabricate and spread fake news.
Formally known as deep synthesis, deep fake refers to highly realistic textual and visual content generated by artificial intelligence.
The new legislation bars users from generating deep fake content on topics already prohibited by existing laws on China’s heavily censored internet. It also outlines take down procedures for content considered false or harmful.
The arrest also came amid a 100-day campaign launched by the internet branch of the Ministry of Public Security in March to crack down on the spread of internet rumors.
Baidu unveiled “Wenxin Yiyan” or “ERNIE Bot” in March. Two months later, Alibaba launched “Tongyi Qianwen,” which roughly translates as seeking truth by asking a thousand questions.
In draft guidelines issued last month to solicit public feedback, China’s cyberspace regulator said generative AI services would be required to undergo security reviews before they can operate.
Service providers will also be required to verify users’ real identities, as well as providing details about the scale and type of data they use, their basic algorithms and other technical information.