China runs ahead of the US on AI regulation

While American leaders fear China could eventually overtake the United States in developing artificial intelligence, Beijing is already far ahead of Washington in issuing rules for the new technology.

Drive the news: Chinese officials will close the consultation on Wednesday on a second round of generative AI regulation, building on a set of rules governing deepfakes agreed in 2022.

Because matter: The Biden administration is behind both allies and adversaries on AI guardrails. While Washington officials talk about granting users rights and urging CEOs to mitigate risks, Beijing and Brussels are effectively offering rights and mitigating risks.

The big picture: If China can lead the way in AI governance, it can project those standards and regulations globally, shaping profitable and flexible markets.

• At the same time, Beijing’s rapid regulation achieves three goals at home:

1. It offers tighter control of the debate by the central government.
2. It builds hybrid business entities that join the Communist Party of China.
3. Increase confidence in AI already among the top tier globally, which drives consumer adoption and spurs growth.

Between the lines: After falling behind the West through consecutive industrial revolutions, Beijing’s leaders are driven by a determination not to be humiliated again in the age of AI.

Beijing thinks clear rules on AI it will help the public trust artificial intelligence.

Yes but: Tech industry insiders see regulation as a hindrance to the rapid development of AI, so many view China’s lead in this area as a US lead.

Being smart: China’s AI regulations regulate what companies in China do with AI, but very few in the West believe they will limit the absolute power of the Chinese government in any real way. US efforts to regulate AI target both businesses and the government.

What are they saying: Matt Sheehan, a China expert on AI governance at the Carnegie Endowment for International Peace, said Axios Beijing’s early regulatory efforts were fueled by fear of AI’s downsides and that it has learned a lot in six years.

• “AI regulation is highly iterative, it’s about building muscle. There’s very little chance you’ll get AI regulation completely right the first time,” Sheehan said.
• After years of struggle, one regulator has emerged as the top AI regulator in China: the Cyberspace Administration of China, known as the CAC.

Details: Authorities have introduced rules every few months since 2021, including a national privacy law, the personal information protection law, and a next-generation AI code of ethics, which outlines requirements that new AI products they offer everything from “enhancement of human-being” to “controllability and credibility”.

• Since 2022, Chinese AI users have transparency rights such as the right to turn off an algorithmic recommendation service, or the right to know when AI-generated content is being served to them.

• Beijing introduced “deep synthesis” rules to guard against deepfakes in 2022. While it didn’t anticipate ChatGPT’s text-based breakthrough, those rules did provide a head start in drafting new rules that address the risks of text-based generative AI .

Where is Xi? “Xi Jinping doesn’t wake up in the morning and set requirements for generative AI models,” Sheehan said. “There’s probably no one on the Politburo in those weeds. They set a high-level direction on wanting to be an AI leader and then they say [to regulators] ‘understand how it’s done’.”

The socket: While Beijing has granted Chinese AI users the rights to, say, contest a hiring algorithm’s rejection of an interview, there’s little chance those rights could be exercised on a large scale, Russell Wald said. , director of policy at the Stanford Center for Human-Centered Artificial Intelligence.

• Chinese corporate and government organizations often lack the ability or inclination to secure individual rights in this way.

Beijing might as well be by regulating now and paying later.

• China’s path to trust in AI comes with problems and burdens for Chinese AI vendors, requiring them to filter content and complete security assessments to maintain control of the Communist Party before new product launches.
• This is the kind of slowdown in innovation that US firms point to as a justification for Washington’s relatively lax regulatory stance.

What’s next: Georgetown-based Toner expects China’s regulations to push its companies to “lean into B2B applications with more limited use cases, as they’re easier to train and use within tight barriers,” and China will struggle to “push forward the frontier of the field”.

The bottom line: China’s top priority is to minimize social disruption while using artificial intelligence. Beijing has also learned from sanctions against Huawei and export controls on US chips that China must be ready to innovate on its own.

• By moving fast on regulation, Beijing is creating a foundation for AI exports across the Global South and to countries participating in its Belt and Road Initiative.

Flashback: China’s privacy law leaves the United States behind