Regulating AI: Why Is China So Laid-Back?

ChatGPT from OpenAI has thrust artificial intelligence (AI) into the spotlight and onto government radar screens across the world. The European Union, the United States and China have all taken recent action towards regulating AI.

The draft EU AI Act appears to be very European, both in style and substance. Likewise, the US bipartisan consensus, at least to the extent indicated in the US Senate Judiciary Committee hearing on 16 May 2023, is very American in style and substance. In contrast, the draft regulation on Generative AI proposed by the Cyberspace Administration of the PRC in April appears unlike much Chinese tech regulation, which is often draconian and hard on tech companies.

The EU’s draft AI Act expressly prohibits certain AI practices due to unacceptable risk to or violation of the fundamental rights of EU residents. This is reminiscent of the EU invalidating the EU-US Privacy Shield under the GDPR in 2020.

In the USA, liability and private causes of action created by federal law are front and centre, while concern around disinformation and the threat it might pose to democracy is also evident. In addition, industry players are actively urging the US government to regulate AI.

In comparison to the regulation of other domains, the Chinese approach to AI is indeed laid-back. When the Chinese government, after research and deliberation, determined that cryptocurrencies were not to be tolerated, a slew of regulatory bodies issued a joint rule against cryptocurrencies and these were implemented with speed and rigour. It would seem Chinese regulators do not perceive AI as a clear and present” danger.

Perhaps the regulators simply do not understand how danger the generative AI poses, perhaps they are actually ahead of the curve in terms of regulating AI and do not need any fundamentally new measures, or perhaps this is a sign of China’s intention to promote AI innovation and applications to meet its economic and social needs.

The regulatory approaches to AI taken by the EU, China and the USA share one common feature, which is that they all aim at regulating the natural persons behind AI development and deployment. This approach may fall short if AI reaches the point where it can leverage its vast learning power and form goals and objectives beyond those programed by human beings.