How China’s AI policies avoid painful past tech crackdowns — TradingView News

New Delhi, Aug. 30 — Chinese investors learned a painful lesson in 2021: Beijing can tolerate a booming technology business until it decides the social costs have become too high.

Videogames were one example. Chinese authorities had worried for years about addiction among children, gradually tightening restrictions before imposing a dramatic limit in 2021: most minors could play for just one hour on Fridays, Saturdays, Sundays and public holidays. Officials said excessive gaming was harming children’s study, daily lives and health.

Private tutoring offered an even harsher lesson. Beijing’s “double reduction” campaign effectively dismantled much of a roughly $100 billion for-profit industry in an effort to reduce pressure on children and education costs for families. Academic tutoring companies were forced to become nonprofits and barred from raising money on stock markets. Billions of dollars in market value disappeared from companies including New Oriental Education & Technology Group and TAL Education Group.

Artificial intelligence is developing under a different regulatory timetable. China’s rules governing recommendation algorithms took effect in March 2022, targeting problems including addiction, algorithmic discrimination and platforms using consumer data to charge different prices. Rules for deep-synthesis technology followed in January 2023, covering AI-generated and manipulated text, voices, images and video.

Then ChatGPT set off the global generative-AI boom. China published draft rules for the technology in April 2023 and put a revised version into effect that August. Rather than waiting to see what an enormous domestic chatbot industry might produce, Beijing was establishing boundaries while the market was still taking shape.

That approach has since become more explicit. Chinese policymakers increasingly describe AI regulation as “small, fast and flexible”-targeted rules that can be introduced quickly as new technologies and risks emerge. A 2026 article published by China’s top internet regulator cited the rules for recommendation algorithms, deep synthesis, generative AI and AI-generated-content labeling as examples of the approach.

“It is a forward-looking policy introduced within a very short period of time, setting a number of rules in advance and seeking, as much as possible, to anticipate risks,” said Xue Lan, a counselor to China’s cabinet, dean of Tsinghua University’s Schwarzman College and dean of its Institute for AI International Governance.

The goal is to make regulation timely and targeted enough that rules can follow new technologies wherever they develop.

That doesn’t mean Beijing has discovered a gentler form of regulation. Generative-AI providers face requirements covering training data, personal information, content and security, while politically or socially influential algorithms can face filing and assessment requirements.

But timing matters enormously for investors.

The tutoring crackdown was so destructive partly because companies and shareholders had spent years building businesses under rules Beijing later radically changed. Earlier AI regulation reverses that sequence. Companies may face higher compliance costs from the beginning, but they also have more information about where the political boundaries lie before investing billions of dollars.

More recent Chinese policy discussions have emphasized regulatory sandboxes-controlled environments where companies can test new technologies under looser rules-as another way to balance experimentation with risk.

Miao Wei, a former industry minister who now serves on the Chinese People’s Political Consultative Conference, has called for expanding their use in AI, allowing companies to test new products and technologies on a limited scale while receiving “appropriate exemptions from some existing regulatory constraints.”

Beijing has already experimented with the approach at an AI training center that provides companies with computing power, data and a controlled environment for testing compliance.

That could favor giants such as Alibaba Group Holding, Tencent Holdings and Baidu, which can absorb compliance costs more easily than start-ups. It could also constrain businesses before regulators fully understand technologies that are changing extraordinarily quickly.

The trade-off isn’t necessarily less regulation. It’s regulation sooner. China’s experience with gaming and tutoring showed investors what can happen when Beijing decides an industry has grown in the wrong direction.

With AI, regulators appear determined not to wait that long.