Last month, we told you about Tencent Music Entertainment‘s integration with DeepSeek.
DeepSeek is the China-headquartered AI platform that triggered concerns across the AI sector earlier this year when its chatbot demonstrated that advanced AI systems can be developed with significantly lower costs and computational resources than previously estimated.
The chatbot’s debut led $157 billion-valued OpenAI to suspect that DeepSeek’s model had been trained on its data without permission, an ironic twist in light of the various copyright infringement lawsuits OpenAI is facing itself. One of those lawsuits was filed by German collection society and licensing body GEMA in November.
TME CEO Ross Liang explained on the company’s earnings call last month that the company had made DeepSeek a key part of its AI music creation tools.
Now, in its Environmental, Social and Governance (ESG) report published today (April 23), the company explains in great detail how its artificial intelligence capabilities have been expanded across the business…
As you will see in TME’s ESG report, which you can read in full here, the company has made substantial investments in AI across sound optimization, audio fidelity, recommendations, music creation and more.
In regards to music creation tools specifically, TME explains that, “as enthusiasm for user-generated content continues to grow, more individuals are eager to participate in music creation”.
According to TME’s ESG report, “by leveraging DeepSeekR1’s AI-powered lyric-generation capabilities alongside QQ Music’s AI songwriting tools, users can craft original songs that resonate with their emotions and personal experiences, further enriching their growth as creative individuals.”
TME says that QQ Music has become the first such platform in China to integrate the DeepSeek model.
Tencent Music Venus
The company (which operates flagship music app QQ Music, as well as Kugou, Kuwo, plus online karaoke service, WeSing) also runs a platform dedicated to AI music creation and production called Tencent Music Venus.
This platform offers several AI-powered music tools, including the AI Tone Magician, a Music Separation application, and the AI Songwriter tool, all of which are “designed to support musicians in their creative processes”.
TME’s ‘AI Songwriter’ tool lets musicians enter “a brief description of a creative idea or custom lyrics” and the tool will then “create original AI-generated songs that include both lyrics and vocals.”
According to TME, those AI-generated songs “can then be easily published on QQ Music with just one click”.
TME works directly with independent artists – and lets them distribute music directly to TME services like QQ Music – via a program called Tencent Musician Platform.
TME reveals in latest ESG report that it counted 580,000 indie acts on the platform as of December 31, 2024.
That number grew by 100,000 in the 12 months to the end of December 2024.
“Users can create original AI-generated songs that include both lyrics and vocals, which can then be easily published on QQ Music with just one click.”
TME ESG report
TME says it provides those indie artists with “comprehensive support across four key dimensions” including “traffic, performances, creation camps, and multifaceted services”.
Those 580,000 indie artists also have access to a range of AI technologies.
TME explains in its ESG report: “We have launched intelligent tools for lyric writing, composition, and arrangement, as well as content-generation tools for short music videos, album covers, and posters.
“These innovations empower musicians in both the creation and promotion of their work. Additionally, we introduced noise-reduction and music-transcription features that have enhanced efficiency in the music creation process.”
TME’s AI Tone Magician technology, meanwhile, allows users to record a voice clip between 30 seconds to 4 minutes to generate a personalized AI-generated vocal tone. After selecting a preferred song, the system can produce an AI-generated version of the track with “a single click”.
TME Music Separation Cloud Application service supports stem separation for uploaded songs, isolating vocal tracks from accompaniment or instrumental layers including drums, bass, guitar, piano, and other mainstream instruments, with the capability to separate up to eight distinct audio tracks.
Music rightsholders are likely to have several concerns regarding TME’s AI tech expansion, particularly as the company develops increasingly sophisticated tools that can generate and manipulate music, which can apparently then be distributed directly to its QQ music streaming platform.
The ESG report doesn’t explicitly detail what music was used to train its models, and rightsholders will likely want to know if their copyrighted works were used without permission or compensation to train these AI systems.
TME’s report says that AI-created content can be “easily published” on QQ, but it doesn’t address questions about attribution/ownership of that content.
That last point is all the more crucial given TME has integrated DeepSeek into its platform. Prominent voices in the industry have recently raised questions about whether DeepSeek itself is training on copyrighted music without permission.
Within TME’s 2024 Annual Report, published alongside the ESG report today (April 23), and which you can read here, TME provides the following breakdown for just how deep the DeepSeek integration is:
“We are incorporating AI into various parts of our product offerings,” states TME.
“For instance, we have integrated the advanced capabilities of DeepSeek large language models (LLMs) to empower users and creators in song creation, creating a more tailored and inclusive experience.
“We also integrated DeepSeek into our AI assistants, comment sections, recommendation pages, and other user touchpoints, allowing users to easily access music information, receive song comments in diverse styles, and enjoy a more personalized music experience.
“Additionally, our AI-powered audio effect analyzes the audio to match songs with the most suitable audio settings to create an optimal audio streaming performance. Furthermore, our AI voice extraction feature allows users to separate original vocals and instrumental tracks with a single click which allows users to sing alongside their favorite artists or enjoy instrumental-only mode for karaoke sessions.”
“Inappropriate or controversial data practices, by us or by others, could limit the acceptance of our AI-enhanced products and content.”
TME Annual report
TME claims in its ESG report that it is “committed to developing and deploying AI technology in a manner that is legal, ethical, and secure.”
The company adds that, “to ensure that relevant Artificial Intelligence Generated Content (“AIGC”) products and technologies comply with legal and regulatory requirements, ethical standards, and the principles of user privacy protection,” it has “enacted and implemented the TME AIGC Product Compliance Manual,” which it says “outlines the overall compliance requirements for AIGC products”.
Within the manual, according to TME, is a section dedicated to “risk alerts and recommendations for cutting-edge AIGC-related issues, such as the appropriate use of AIGC training data, the ownership and copyrightability of AIGC, and AIGC patent and trademark protection”.
In TME’s Annual Report, the company acknowledges that it is “subject to risks and uncertainties associated with the evolving AI regulations, potential infringement claims and increasing costs for regulatory compliance. ”
TME explains: “As with most emerging technologies, AI comes with its own set of risks and challenges that could affect its adoption and our business. AI algorithms may be flawed, and the data used could be incomplete or biased. Inappropriate or controversial data practices, by us or by others, could limit the acceptance of our AI-enhanced products and content.
“Furthermore, there are uncertainties around the ownership and intellectual property protection of AIGC products. Using AIGC tools could also lead to potential copyright infringement and other legal challenges. If we are unable to secure the needed permissions or licenses for using AI tools-whether because we cannot identify the rights holder or for any other reason-we might infringe on others’ rights which could lead to monetary claims, fines, penalties, or less content for our users.
“The regulatory and legal framework on generative AI is evolving rapidly and might not fully address every aspect of its research, development, and use.”
In the very next risk section within the report, dedicated to “Pending or future litigation”, Tencent Music reveals that, as of December 31, 2024, there were 315 lawsuits pending in connection with alleged copyright infringement on its platform, with an aggregate amount of damages sought of approximately RMB 198.7 million (USD $27.2 million).
It does not specify whether or not any of those infringement claims were due to AI-generated content.
Tencent Music Entertainment, China’s largest music streaming company, saw its annual subscription revenue surge 25.9% YoY in 2024 to RMB 15.23 billion (USD $2.12 billion at the average exchange rate for 2024) from RMB 12.1 billion in 2023 (USD $1.71 billion at the average exchange rate for 2023).
The company attributed this performance to the expansion of its paying user base and improved average revenue per paying user (ARPPU) boosted by its ‘Super VIP’ (SVIP) tier.
TME’s paying subscribers increased 13.4% YoY to 121 million in Q4 2024 from 106.7 million in Q4 2023. This also reflects an addition of 2 million subscribers between Q3 and Q4 2024. ARPPU rose to RMB 11.10 ($1.54) in Q4 from RMB 10.70 ($1.49) in Q4 2023.
The invasion of AI music on streaming services was highlighted last week when France-headquartered music streaming service Deezer revealed that approximately 18% of all tracks now being uploaded to its platform are fully generated by artificial intelligence.
The company disclosed last Wednesday (April 16) that more than 20,000 AI-generated tracks are being delivered to its platform every day – around double the 10,000 daily AI uploads Deezer reported in January.
“AI-generated content continues to flood streaming platforms like Deezer and we see no sign of it slowing down,” said Aurelien Herault, Chief Innovation Officer at Deezer.Music ComeOn
