Last week, I appeared on the CBC News podcast “Nothing Is Foreign” about the draft regulations and what it means for the Chinese government to act so quickly on a still very new technology. I talked about whether
As I said in my podcast, I see the draft regulation as a mixture of judicious limits on AI risks and a continuation of the Chinese government’s strong tradition of proactive intervention in the tech industry.
Many of the draft regulation’s provisions are principles asserted by AI critics in the West: Data used to train generative AI models must not violate intellectual property or privacy. Algorithms should not discriminate against users based on race, ethnicity, age, gender, or any other attribute. AI companies should be transparent about how they obtained their training data and how they hired humans to label the data.
At the same time, there are rules that other countries would probably hesitate to follow. The government, like other social platforms in China, requires people using these generative AI tools to register their true identities. Content generated by AI software should also “reflect socialist core values.”
None of these requirements should come as a surprise. The Chinese government has heavily regulated tech companies in recent years, penalizing loosely regulated platforms and incorporating new products into established censorship regimes.
This document makes the regulatory traditions clearer. Other regulations passed in China, such as personal data, algorithms, deepfakes and cybersecurity, are also frequently mentioned. In a way, it feels like these separate documents are slowly forming a framework. A web of rules to help governments meet the new challenges of the technology age.
The fact that the Chinese government can react so quickly to new technological phenomena is a double-edged sword. Matt Sheehan, a fellow at the Carnegie Endowment for International Peace, writes that the strength of this approach, which considers every new technology trend in isolation, is “the precision that yields concrete remedies for specific problems.” “The weakness is its piecemeal nature, requiring regulators to develop new regulations for new uses and issues.” You may miss an opportunity to think strategically. You can contrast this approach with that of the EU, which, as my colleague Melissa recently described, has been working for years on a “very ambitious” AI law. (Recent revisions to the AI law include regulations on generative AI.)
