China starts strict laws for AI-generated content, targets social media influencers

China has introduced strict regulations requiring mandatory labelling of AI-generated content. It’s a part of the Chinese government’s plan to tackle rising concerns over misinformation, fraud and copyright issues, the South China Morning Post reported.

Users must now declare if the content is AI-generated. At the same time, service providers must retain records of such content for at least six months. Tampering with or removing AI labels is strictly prohibited. Violations will cause penalties.

These measures are part of China’s effort to tighten control over its digital space. The Cyberspace Administration of China (CAC) has made AI regulation a major focus of its 2025 “Qinglang” (Clear and Bright) campaign, which aims to clean up the internet.

The campaign targets the spread of false information, manipulative content and misuse of AI. It targets “Internet water armies”, social media influencers who are paid to sway public opinion.

Other goals include monitoring short-video platforms, curbing deceptive influencer marketing and protecting underage users online. Support for mandatory labelling is growing, especially with the rise of local AI models like DeepSeek, Qwen (by Alibaba) and Manus by start-up Butterfly Effect, the publication added.

Globally, countries are taking similar steps. The EU’s AI Act mandates AI content labelling. The United States and the UK are working on legislation focused on transparency and compliance.

However, experts warn that labelling alone may not be enough. Challenges include regulating real-time AI applications like live streams and voice calls. Watermarks and metadata can be altered or removed easily, they warn. Inconsistent detection methods across platforms complicate enforcement.

AI regulation in India

India has no specific AI laws yet, but it has launched key frameworks like the National Strategy for AI (2018), Principles for Responsible AI (2021) and Operationalising Principles for Responsible AI. The idea is to guide ethical, transparent and accountable AI development.