In a recent crackdown, Chinaโs internet watchdog named and shamed agents caught offering unauthorised local access to ChatGPT. The US provider OpenAI has blocked the use of its artificial intelligence (AI) technologies on the mainland, Hong Kong, and Macau.
Crackdown on unauthorised access
The Cyberspace Administration of China (CAC) has punished several website operators for offering unauthorised access to generative AI (GenAI) services, including ChatGPT. This action highlights China’s resolve to enforce domestic AI rules, introduced in August last year, which require all AI services to be screened and registered before public availability. Some developers and enterprises have used virtual private networks to bypass these rules and access unregistered GenAI services.
The crackdown coincides with OpenAI’s move to block access to its application programming interface (API) from July 9 in “unsupported countries and territories.” This ban covers the mainland, Hong Kong, Macau, and US-sanctioned states like Iran, North Korea, and Russia.
Nanchuan District Rongcheng Network Technology Studio, operating over ten websites in China, was singled out by the CACโs branch in Chongqing to provide ChatGPT services without undergoing a safety assessment. The CAC’s Chongqing branch summoned the studio’s representatives and ordered them to disable access to ChatGPT immediately.
Punishments and enforcement
The operators of Yizigpt.com, an AI writing platform; Kukupao.com.cn, a gaming site; and Lvshifuwuwang, a legal services provider, were also caught for failing to censor content containing illegal information generated from undisclosed GenAI services. These operators faced “administrative punishment,” including fines, business shutdowns, or detention of responsible individuals.
Authorities in other mainland cities have also taken action against the inappropriate use of AI. In Changsha, the capital of Hunan Province, police detained a man named Xu for three days after being caught with an AI-generated fake reward notice for wanted suspects. Similarly, in Leiyang, Hunan, a man surnamed Zheng was caught generating fake information via AI and posting it on a local website. Both Zheng and the website operator faced administrative punishment.
Beijingโs stance on GenAI
Since the launch of ChatGPT in November 2022, Beijing has maintained a cautious stance on GenAI. While the mainland aims to become a world leader in AI, it also wants to ensure that GenAI service providers “adhere to core socialist values” and do not generate content that “incites subversion of state power and the overthrow of the socialist system, endangers national security and interests,” as per the GenAI regulation effective last year.
So far, 117 GenAI products have been registered with the CAC. In January, 14 large language models and enterprise applications were approved by authorities for commercial use, following the initial approval of GenAI services released to the public last August.
