Chinese regulators have ordered local tech giants not to provide access to ChatGPT due to the inability to censor responses. Nikkei Asia writes about it.
According to “people with direct knowledge of the matter,” the authorities instructed Tencent and Ant Group not only to restrict access to the chatbot, but also to report the launch of their own services.
Although ChatGPT is not officially available in China, it has caused a stir among the country’s Internet users and the AI community.
The latter expressed concern that the technology was not developed in China. Some of the experts cited China’s strict technical regulation and censorship as barriers to the creation of such chatbots.
According to Nikkei, Chinese users were able to access ChatGPT through VPN services or third-party messaging app integrations. Earlier, Tencent, the developer of the popular WeChat messenger, blocked some attempts to introduce a chatbot into the program.
At the same time, China Daily, China’s largest English-language newspaper, warned that OpenAI’s chatbot could be used to spread Western propaganda.
#ChatGPT can potentially amplify US disinformation campaign. An expert points out only a few articles can influence the output of ChatGPT on specific questions. #MediaUnlocked https://t.co/QaYqjSjx3U pic.twitter.com/7bh9icVQvm
— China Daily (@ChinaDaily) February 20, 2023
Another Chinese reporter, Xu-Pan Yu, posted a video asking a chatbot about Xinjiang. In response, the system cited “reports of human rights abuses against Uyghur Muslims, including mass incarceration in ‘re-education’ camps, forced labor and other forms of persecution by the Chinese government.”
According to the journalist, the generated text is fully consistent with the theses of the United States.
Representatives of the Chinese technology industry said that the ban of regulators did not come as a surprise.
In February, the Beijing authorities announced their intention to support local chatbot developers.
In January, Chinese regulators tightened regulations on the production of deepfakes.
In September 2022, journalists reported that the Chinese authorities politically censored Baidu’s ERNIE-ViLG text-based image generator.