Is this singer real? Why China wants to regulate AI-generated humans

CHONGQING - Clad in a stylish black sweatshirt, Chinese singer Pin is seen in a video strolling along a street as he introduces one of his songs, a cover of Love You Very Much by Taiwanese singer Rene Liu.

The video, posted on April 3 on his supposedly official Instagram account, drew questions from followers.

“I’ve always thought he was AI,” one of them wrote, clearly puzzled.

To which another replied: “You can’t possibly believe this is real, right? It’s obviously AI.”

Yet another said: “How can you tell the difference when they look so real now?”

The real-or-not question poses potential problems in China’s fast-growing “digital humans” market, valued at 4.12 billion yuan (S$769 million) in 2024 and expected to reach 25 billion yuan by 2029.

Potential problems range from users being misled to have delusions or being encouraged to harm themselves, to the erosion of trust in the technology or the larger society.

The authorities hope to address these issues with new laws slated to take effect later in 2026, including mandatory declaration of digital humans as such.

As China sets its sights on becoming a global leader in artificial intelligence amid an intensifying tech and trade war with the US, the laws are meant to provide companies with clarity on how to develop their AI technologies and protect users at the same time.

Digital humans refer to hyper-realistic virtual beings created using AI. They are modelled to speak, behave and generally interact with people in real time. They may also be based on real people, from celebrities to politicians to athletes. 

Such virtual beings are being used in various industries in China to provide services such as museum tours, news broadcasting, live streaming and virtual romantic relationships.

There are also companies recreating in virtual form people who have died for their loved ones who miss them.

China’s state news agency Xinhua released the world’s first AI news anchor in 2018.

In June 2025, an AI clone of influencer Luo Yonghao held an online sales session that raked in 55 million yuan in about seven hours.

When AI creations are declared as such, there is amazement all around and across the world.

When they are not labelled as AI or even merely not clearly human, head-spinning controversies can arise, such as in the case of singer Pin, who is also known as Datouzhen.

Is Pin a human singer using an avatar? Or is Pin an AI copy of an AI singer – or maybe even an AI copy of another AI copy of an AI singer?

On his Instagram account, he is described as a “rising star” in the Chinese music scene who is “born in 1999” and gained widespread attention online in late 2024 for his husky, emotive vocals and rock interpretations of ballads such as Look, Look At The Face Of The Moon.

Pin’s “Datouzhen” handle and videos match that of an AI-generated online singer, also known as Datouzhen, who started releasing covers in late 2025 on Douyin, TikTok’s China counterpart.

To add to the confusion, there is an earlier Douyin account, also named Datouzhen, which started releasing music covers in late 2024. Said to be 35 years old, he had gained widespread attention in China after attracting 800,000 followers in two months, according to media reports.

Netizens have asked whether this Datouzhen, who does not show his face, is indeed a human being or AI-generated, but they have not received a satisfactory answer so far.

Labels for digital humans

China is pushing to implement laws on digital humans as part of its drive to use advanced technology to power its economy.

The proposed new rules include a requirement for digital humans to be labelled properly as AI content, a ban on the use of other people’s personal information to create digital humans without their consent, and having providers take necessary measures to intervene and provide professional assistance if users show tendencies of self-harm.

These draft laws – currently titled Provisional Measures on the Administration of Human-like Interactive AI services – were opened for public feedback earlier in April.

The 32-article regulation also spans areas such as the protection of users’ mental health by having the system generate consolation or encouragement to seek help, to the stipulation that elderly users set emergency contacts to prevent them from being scammed.

Minors, too, will be better protected, as laws will require providers of AI services to restrict the amount of time they spend online and the content they can watch, among other provisions.

Experts told The Straits Times that the proposed laws are a crucial step towards addressing dangers that have become increasingly obvious with the rapid expansion of the use of digital humans. They include deep addiction among the youth, social alienation and cognitive manipulation.

Associate Professor Jiang Tianjiao at Fudan University in Shanghai said the changes “establish an explicit safety assessment system and technical requirements for privacy protection and mental health safeguards”.

But experts say questions remain over the practicality of the requirements.

Tech consultancy Concordia AI noted on media platform Substack in February that “there is some uncertainty whether the proposed approaches, such as manual intervention in high-risk situations, are feasible at scale”, based on an earlier version of the proposed laws.

The upcoming laws are expected to be enforced by the Cyberspace Administration of China. But there is no clarity yet on penalties if lines are crossed.

On a more positive note, the current proposed rules have laid the foundation for innovation in the industry and are expected to give China a leg-up in its rivalry with the US.

Adjunct Associate Professor Adam Au, who teaches AI, law and data protection part-time at the University of Hong Kong, said that the proposed laws help to formalise the industry. This will allow companies to plan and know what they have to disclose, while investors “have a clearer sense of the risks and enforcement tools on the table”.

“In a system that prizes stability, this kind of legal bundling is often the precondition for scale, especially in China,” he added. 

The new laws also make it clear that the regulators are serious about protecting humans, whether they are consumers or models used to create digital beings.

This means companies cannot use digital humans based on certain celebrities without their consent, as some have done. In November 2025, Chinese actress Wen Zhengrong became a victim of such theft when multiple live streams used AI-generated deepfakes of her to sell products.

Instead, companies have to look elsewhere for inspiration, or perhaps pay commission to artiste management agencies.

The rule about obtaining consent from people on whom digital humans are based sends the message that “spinning up a ‘virtual you’ without permission is a clear rights violation, not clever entrepreneurship”, said Prof Au.

“On the surface, the language (of the laws) leans in favour of users (with) more talk of consent, privacy, minors and ‘healthy’ content than in some earlier AI rules.

“Underneath, the structure is growth‑compatible. It brings a sprawling sector inside a familiar governance perimeter so that it can expand without triggering a crisis of confidence,” he said.

For fans of Datouzhen, the question of whether their idol is human or not can be finally put to rest.