China’s leadership wants to embrace AI advances but also control risks, as ChatGPT shocks with power and popularity

[Hamartia Antidote]

China to embrace AI advances but also control risks as ChatGPT wave spreads

Politburo recently concluded that China must ‘pay attention to the development of AGI’ but also ‘take risk prevention into account’.

www.scmp.com

• It appears that while advancement of AI is desired by Beijing, there are also concerns about the consequences of such actions
• Politburo recently concluded that China must ‘pay attention to the development of AGI’ but also ‘take risk prevention into account’

Politburo recently concluded that China must ‘pay attention to the development of AGI’ but also ‘take risk prevention into account’. Photo: Shutterstock

China’s leaders have gone public with their concerns about the potential risks from artificial general intelligence (AGI), a more powerful version of current AI technologies, as Microsoft-backed OpenAI’s ChatGPT product continues to take the global tech community by storm.

In a statement issued by state media outlet Xinhua on Friday summarising the Politburo’s quarterly meeting on China’s social and economic development, the Politburo – headed by President Xi Jinping – concluded that China must “pay attention to the development of AGI, create an ecosystem for innovation but at the same time take risk prevention into account”.

This is the first time that Beijing has included AGI within a Politburo statement, reflecting the country’s mixed signals to date on where it stands on ChatGPT-style services. It appears that while advancement of AI is desired, there are also concerns about the consequences of such actions.

ChatGPT is not available for China-based users, and China’s internet regulator has already published regulations trying to restrict AI-generated content to within official Chinese narratives.

The Cyberspace Administration of China, the internet watchdog, earlier this month drafted new rules demanding generative AI developers pass over their products for a security assessment before being made available to the public. This would be to avoid content being generated that contravenes rules relating to discussion of regime subversion, violence or pornography.

At the same time, the AI frenzy triggered by ChatGPT, which was launched last November, has also prodded Chinese tech giants to catch up in the space. Search engine giant Baidu, e-commerce behemoth Alibaba Group Holding, video gaming company NetEase and AI firm SenseTime have all launched their own ChatGPT-like services. Alibaba owns the Post.

With regulators circling the new technology, Chinese Big Tech firms have been careful not to step out of line until policy is clear.

A review of four Chinese chatbot services including Baidu’s Ernie Bot, Alibaba’s Tongyi Qianwen, Spark Desk by iFlyTek and SenseTime’s SenseChat, showed that all of their bot’s terms of use stipulate that early testers should not use their services to spread information that “harms national security” or “subverts state regime” and must “maintain a ‘clean and clear’ online environment”.

Analysts say that Chinese authorities are trying to balance innovation with control.

China’s regulations “give the government broad authority to restrict activities that endanger national security interests, damage China’s national image, or disrupt the economy”, said Hanna Dohmen, research analyst at Georgetown’s Center for Security and Emerging Technology, in February.

“The regulators also face social and institutional pressures, and they need to show that they are carrying out responsive regulation on all these emerging issues,” said Dr You Chuanman, director of the IIA Centre for Regulation and Global Governance with the Chinese University of Hong Kong, Shenzhen Campus. “They have to take action one way or another, whether it’s welcomed by the market or not.”

You said despite concerns, China is still promoting infrastructural research on AI but the tricky part is when it comes to the application of the technology that involves consumer access to uncensored information.
“In essence, it’s about the content [generative AI produces] which falls under the most debated issue of Chinese internet regulation,” he said.