China takes comprehensive measures to curb AI misuse

Artificial intelligence (AI), deep synthesis, and other emerging technologies are developing rapidly, playing an increasingly important role in fostering new quality productive forces and driving high-quality economic and social development, while also bringing great convenience to everyday life.

However, these technologies have also been misused to spread disinformation and disrupt the online environment, causing negative impact.

In response, China has taken a comprehensive approach to curb the misuse of AI and promote its responsible development that serves the common interests of society. In September this year, China released a set of guidelines on labeling internet content that is generated or composed by AI technology. These rules mandate clear labeling to help users identify fabricated information, define the labeling obligations of service providers, and regulate the labeling process throughout content creation and dissemination.

Addressing impersonation and fraud

"Hey everyone, I've got a special favor to ask today… I'm offering discounts on 300 crates of fresh local eggs from my hometown!"

In one recent case, a widely shared short video appeared to show a well-known athlete promoting eggs from their hometown, drawing enthusiastic support from fans. Yet, the video, like others from the same account, was AI-generated. One product link associated with the video showed more than 47,000 units sold.

According to a representative from a major short-video platform, impersonation of public figures through AI typically takes several forms: using AI-generated images of well-known individuals as profile photos, producing unlabeled AI-generated videos featuring celebrities, and creating fake celebrity accounts to attract followers and generate profit through deceptive content.

China's Civil Code and other relevant regulations already provide protection for portrait and voice rights. The new labeling regulations further reinforce these protections by requiring platforms to clearly mark AI-generated content.

"Short video platforms must not only fulfill their own responsibilities but also collaborate with users to strengthen co-governance," said an executive of Chinese tech firm Tencent. When users encounter unlabeled AI-generated short videos, they file a complaint with relevant materials. Once the complaint is verified, the platform will add visible floating labels to the content.

Douyin, the Chinese version of TikTok, has introduced a portrait protection database for well-known figures to better detect impersonation and fraudulent marketing content.

Combating fake news and strengthening oversight

In another incident, a fabricated "video report" of a large fire caused public panic. A subsequent investigation found that the account responsible used AI to mass-produce disinformation, generating between 4,000 and 7,000 fabricated news items per day and earning more than 10,000 yuan ($1,413.32) daily.

Despite efforts to combat such abuse, gaps remain. According to the head of content safety at a news platform, anti-disinformation mechanisms are not always timely because AI-generated images or videos may go unflagged for hours, limiting early warnings to the public.

Yang Qingwang, deputy dean of the School of Law at Central South University, emphasized the need for platforms to respond quickly, using authoritative information to flag and remove false content and penalize violating accounts.

In recent years, China has steadily strengthened its legal and regulatory framework to define the boundaries of responsible AI use. Meanwhile, regulators nationwide are exploring new approaches to counter the spread of AI-generated fake news.

"We work closely with research institutions and technology enterprises to strengthen monitoring, early detection, and identification of malicious deepfakes. Through collaboration with relevant authorities, we have established a closed-loop mechanism covering rumor inspection, verification, joint assessment, case transfer, and law enforcement, allowing for a timely and legal response to AI-generated disinformation," said an official from the cybersecurity division of the Shenyang public security bureau in northeast China's Liaoning province.

Building digital literacy and defenses

Platform operators also encourage users to remain vigilant. Signs of AI-generated content may include logical inconsistencies, flat or overly uniform emotional tone, or vague factual details, as well as unnatural facial movements, irregular lighting, or distorted audio. Users are advised to verify sources and rely on authoritative media outlets or official channels.

Through coordinated efforts among regulators, platforms, and users, and supported by technological detection, reporting mechanisms, and legal accountability, China aims to build a robust defense against AI-generated misinformation and ensure the healthy development of emerging technologies.