AI voice cloning used for unauthorized commercial speech and fraud in China
Voice actor Ma Zhengyang alleges his voice was cloned without permission and used in a drama's advertisement and animation. Reports indicate AI voice cloning software is widely available in China and being misused for commercial purposes and telecommunications fraud targeting elderly people.
What happened
Voice actor Ma Zhengyang reported in August 2026 that his voice was cloned without authorization using AI technology and used in a pharmaceutical advertisement inserted into the drama 'Jiumen' and in animation sequences of the drama.
According to a Xinhua commentary, AI voice cloning has become widespread in China, with voice samples being extracted from individuals without permission, commercialized in short videos and product promotion advertisements, and sometimes sold openly on online shopping and second-hand trading platforms.
Criminals have begun using AI-cloned voices in telecommunications fraud schemes, impersonating family members and relatives to trick elderly individuals into transferring savings by fabricating emergencies such as traffic accidents and medical expenses.
Reported impact
Individual voice rights have been violated through unauthorized commercial use of cloned voices, and victims face difficulties in identifying and proving infringement due to complex distribution paths and multiple actors in the technical chain.
Elderly people have been defrauded by AI voice cloning used to impersonate family members in emergency scams, resulting in loss of retirement savings.
Index analysis
AI voice cloning technology has low barrier to entry, with software priced from ten yuan to hundreds of yuan and available on multiple platforms, with some shops recording thousands of transactions.
Technical responsibility is fragmented across data providers, model developers, and platform operators, making it difficult for victims to determine whom to sue and how to gather evidence.
Governance controls
Platforms should upgrade audio authentication systems and remove illegal voice model sales. Judicial authorities need to clarify standards for recognizing voice rights and increase penalties for infringement and fraud.
Regulatory authorities should require AI training platforms to conduct compliance reviews of data sources, and content distribution platforms should establish rapid-response mechanisms for handling infringement complaints.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.