Китай усиливает контроль за человекоподобным ИИ для повышения безопасности и этики Translation: China strengthens control over human-like AI to enhance safety and ethics.

The Chinese cybersecurity regulator has released a draft of rules for public discussion regarding the supervision of AI services that mimic human personalities. This was reported by Reuters.

The regulations will pertain to products and services based on artificial intelligence that emulate human-like thought processes and communication styles, engaging with users on an emotional level. These interactions may include text, images, audio, video, and other forms of media.

The document aims to reinforce requirements for providers of such services in terms of safety and ethical standards.

The draft rules mandate that service providers:

The proposed measures also prohibit AI services from generating content that threatens the national security of the PRC, spreads rumors, or promotes violence and obscenity.

As far back as 2021, the Ministry of Science and Technology in China introduced a set of ethical principles for AI regulation focused on protecting user rights and mitigating risks. Key points included ensuring the «manageability and reliability» of AI, fairness and justice, and enhancing ethical literacy.

Two years later, the Cyberspace Administration published rules for regulating generative AI algorithms. Among the requirements such as ensuring the reliability of content and respecting intellectual property, there was a specific mention of prohibiting calls to undermine the government.

It is worth noting that Chinese authorities identified a risk of overheating in the market for humanoid robots.