China has issued a set of draft regulations that will require chatbots using artificial intelligence and developed by its tech giants to stick to the ruling Communist Party line, amid comments that the move will likely be the death of further innovation in the sector.
“The content generated by generative artificial intelligence should reflect the core values of socialism,” the country’s Cyberspace Administration said in draft rules issued for public feedback and comment on April 11.
“[It] must not contain subversion of state power, overthrow of the socialist system, incitement to split the country, undermine national unity [or] promote terrorism [and] extremism,” it said, using phrases typically used to target public dissent and criticism of the government.
“Content that may disrupt economic and social order” is also banned, according to the draft rules, which the government wants to see implemented by the end of the year.
The rules come as Chinese tech firms rush to launch homegrown AI chatbots, amid reports that regulators have warned major tech companies not to offer the Microsoft-backed artificial intelligence bot ChatGPT to the public.
They reflect official concerns around any technology that can produce content without the prior approval of government censors.
In 2017, Tencent took down its chatbot Baby Q after it referred to the government as a “corrupt regime,” claimed it had no love for the Communist Party and said it dreamed of emigrating to the United States, amid reports that its programmers had been hauled in for questioning by police.
Nonetheless, the rules also claim that the government supports “independent innovation, popularization and application” in technologies like artificial intelligence algorithms and frameworks.
However, organizations and individuals using AI products to provide services will be held responsible for their output, it warns.
Social media comments said the rules would likely sound the death knell for AI innovation in the sector, which has already seen a demo version of Baidu’s homegrown chatbot Ernie launched with only a preselected set of questions allowed at the press conference in mid-March.
“This subversive AI is hereby sentenced to death. Sentence to be carried out immediately,” quipped Twitter user @Meta_epoch, while @TTL0001 wondered if the bot would need to fill out an application for Communist Party membership.
According to user @linglingfa, the rules mean that this style of AI “will definitely be banned in China now.”
“AI that has to encompass sensitive words is basically dead in the water,” agreed user @hunterpig586, in a reference to the list of politically sensitive terms currently banned from China’s internet.
“The AI should also study the 20 Principles, support the Two Establishes, consciously uphold the Two Maintains and achieve the Four Self-confidences!” quipped Twitter user @Alexajinyu in a sarcastic reference to key elements of supreme leader Xi Jinping’s personal ideology.
“This is ridiculous — why don’t they just give up altogether, really,” added @luhaha777, while @Kev1nLee214 added: “It’s clear that [AI chatbots] are destined to become a tool for ideological output … as their input material is ideological in the first place.”
And user @swift_pink1231 wondered: “What will they do when AI evolves further and is able to block them out?”
Australia-based computer scientist Zhang Xiaogang said the censorship of chatbots would likely kill innovation in the sector.
“A dictatorial regime will always try to control everything, but this is a ridiculous approach,” Zhang said. “Restricting such things is tantamount to restricting AI itself, which will cause China’s AI to fall behind the rest of the world.”
“All China will be able to do then is steal other people’s technology,” he said.
Social media influencer Great Firewall Frog, who uses the Twitter account @GFWFrog, said the first priority of Chinese development is to curb and suppress innovation.
“The first priority of AI research teams is … to comply with the relevant laws and regulations,” he said. “In China in the new era under Xi Jinping, the Communist Party spends most of its AI research efforts trying to find and filter out content that it deems insulting to China, and to maintain the stability of the regime.”
“The Communist Party’s AI can only become even more focused on the ruling party,” he said, adding sarcastically: “If it’s not careful, it could get accused of picking quarrels and stirring up trouble, or incitement to subvert state power.”
Veteran economist Li Hengqing said the new rules show that the ruling party will have to keep tightening controls on public speech if it is to stay in power in the long term.
“The important thing is the content, the so-called ideology,” Li said. “Ideology is China’s current weakness.”
Media commentator Wang Jian said Beijing only has a use for innovation if it helps maintain the party’s grip on power. “The Chinese Communist Party’s core value is to stay in power,” he said. “To do this, it has two tools at its disposal: the gun and the pen, and the pen is where ideology comes in.”
“The Communist Party would prefer not to allow an industry to develop if it could threaten its ideological controls,” he said.