Tencent Music Removed 250,000 Songs & Reviewed 600,000 Copyright Cases in 2025. The AI Problem Is Getting Harder.
China’s largest music streaming company is fighting the same battle as every other major platform in the world, just at a different scale. Tencent Music Entertainment disclosed in its 2025 Environmental, Social and Governance report, published in April, that it removed more than 250,000 policy-violating songs from its platforms last year and reviewed over 600,000 cases involving what it classified as “high-risk copyright content.” The figures cover QQ Music and TME’s other streaming services and were generated through a combination of AI-powered detection tools and manual inspection. The company flagged the removed content as posing reputational risks or violating platform content policies.
The 250,000 figure is the headline number, but a separate disclosure within the same report is arguably more telling about where platform-level copyright enforcement is heading. TME said it removed over 27,000 songs specifically categorized under three distinct forms of gray-market manipulation: “song theft,” defined as tampering with rights ownership information; “song laundering,” defined as plagiarism or alteration of existing music works; and “trend hijacking,” defined as speculative or misleading content creation designed to game distribution systems. The company described these practices as “becoming increasingly covert and complex,” language that reflects a sophisticated adversarial dynamic between platform enforcement and bad actors who adapt their methods in response to detection. To combat these issues, TME has deployed multimodal recognition technologies including audio fingerprinting, voiceprint recognition, melody comparison, and text similarity analysis for continuous catalog monitoring.
The governance infrastructure TME has built around these enforcement activities is notably structured. The company released version 2.0 of its Code of Conduct for Inappropriate Content Management in 2025, introducing a tiered control mechanism with scoring standards and graduated response measures. It also established a “health rating” system for labels that continuously assesses their compliance performance, with consequences ranging from warnings and mandated rectifications to suspension or termination of collaboration for those that fail to meet standards. That label accountability layer is a meaningful addition: it shifts some enforcement responsibility upstream to the distributors and rights holders rather than relying entirely on platform-level detection after the fact.
Related Stories
On AI-generated content specifically, TME’s approach mirrors the detection-first model that Deezer has deployed most aggressively among Western platforms. The company said it now identifies and tags content containing AIGC using audio models and implements governance actions based on its internal rules, and has established a regular content security assessment and detection mechanism for AI-generated material. On the question of AI model training, which has been the central point of conflict in Western label litigation against Suno and Udio, TME stated it adheres to the principle of “legally compliant licensing” and only incorporates content into AI model training “after obtaining licensing from copyright holders.” The company also said it has introduced AI-specific clauses in product user service agreements to clarify responsibilities regarding content usage and authorization.
The contrast with the Western market is instructive. In the US and Europe, platform-level AI transparency is being built through a mix of voluntary self-disclosure systems like Spotify’s AI Credits beta and Apple Music’s Transparency Tags, alongside active detection tools deployed by Deezer. The legal framework underlying those systems is still being contested in court. TME operates in a more centralized regulatory environment where the Music Copyright Society of China, with which it signed a strategic copyright cooperation agreement in 2025, provides a more unified licensing and enforcement infrastructure. TME also co-initiated the establishment of an Artificial Intelligence Copyright Working Committee under China’s Copyright Society, where it serves as deputy director, and took the lead in drafting cybersecurity standard practice guidelines for AI-generated audio content labeling. That standard-setting role gives TME a level of structural influence over how AI content is governed in China that no single Western platform currently holds.
The scale of TME’s own copyright exposure adds necessary context to these enforcement disclosures. According to the company’s 2025 annual report filed with the US SEC, 160 lawsuits were pending against it for alleged copyright infringement on its platforms as of December 31, 2025, with aggregate damages sought of approximately 187.1 million Chinese yuan, or roughly $26.8 million. TME said it does not believe any such proceedings are likely to have a material adverse effect on the company, a standard legal qualifier that nonetheless illustrates the volume of rights disputes any platform of this scale manages simultaneously. As the broader music industry continues to grapple with AI content flooding distribution pipelines, the TME report offers one of the most detailed public accounts yet of what platform-level enforcement at scale actually looks like in practice.
Related Stories
Related Stories
Related Stories
Related Stories
FTM Newsletter
Sign Up for the Weekly Flare Newsletter so they news comes to you!
