The special action of "Clear - Rectify AI Technology Abuse" is a special action deployed nationwide by the Central Internet Information Office (CNIO).
In April 2025, the Central Internet Information Office (CNIO) issued a notice to deploy a three-month nationwide special action called "Clearing the Air - Rectifying the Abuse of AI Technology".
The special operation is divided into two phases, focusing on AI technology governance and illegal content remediation:
Phase 1: Strengthening AI technology source governance
The focus is on the following six categories:
Illegal AI products: not filed, providing illegal functions (e.g. "one-click undressing"), violating others' privacy (e.g. cloning voice and face).
Sale of offending tutorials and tools: dissemination of tutorials for faking face and voice changes, sale of voice synthesizers, face-changing tools, and so on.
The training corpus is not standardized: data from infringing or illegal sources are used and no management mechanism is in place.
Weak security measures: Lack of audit mechanism and management of violating accounts, and ineffective supervision of API access services by the platform.
Failure to label synthesized content: Failure to label in-depth synthesized content, leading the public to misjudge the authenticity of the information.
Risks in key areas: AI in sensitive areas such as healthcare and finance have no professional security audits and misleading services.
Phase II: Remediation of AI production and dissemination of illegal and undesirable information
The focus is on the following seven categories:
AI Rumor-mongering: fabricating rumors about current affairs, social hotspots, disaster events, etc., and releasing them under the guise of official releases.
Publishing inaccurate information: editing and piecing together false content, rehashing old news, spreading pseudoscience.
Generate pornographic and vulgar content: AI stripping, soft porn images, horror images, small yellow text, etc.
Impersonation of others: AI face-swapping, voice cloning to impersonate public figures for fraudulent purposes, spoofing the deceased.
Engaged in network water arming activities: AI batch operating accounts, brushing traffic control comments, creating false heat.
Non-compliant AI apps and services: fake websites, vulgar chats, non-compliant amplification tools, etc.
Infringement of the rights and interests of minors: inducing indulgence, dissemination of undesirable content in the mode of minors.
Goal: Through this special action, we will comprehensively improve the governance of AI technology and content, crack down on illegal use, and protect the legitimate rights and interests of the public.
