SoundOn, the music distribution and promotion platform owned by TikTok, is cracking down on unauthorized uploads by deploying a new detection service from content recognition firm ACRCloud.
The partnership will integrate ACRCloud’s newly developed Derivative Works Detection tool to detect copyrighted works that have been uploaded via SoundOn without authorization.
The technology can detect these tracks even when they have been “significantly altered” by speed or pitch shifting, according to an announcement on Thursday (April 2).
The tech uses audio fingerprinting to identify these modified works before they reach streaming platforms, flagging them for pre- and post-distribution review.
With Derivative Works Detection, SoundOn will use what it describes as a “rigorous customer identification process,” requiring uploaders to verify their identity with photo ID, with flagged content escalated to “human review”.
TikTok said that the partnership adds “an extra layer of anti-fraud detection” into SoundOn.
Paired with TikTok’s internal content scanning infrastructure, the company says the combined system creates “a best-in-class, industry-leading, multi-signal detection framework” that reduces the risk of unauthorized uploads.
SoundOn says this ensures that the content delivered to DSPs like Spotify, Apple Music and Amazon Music is “original, authorized, and trusted.”
“IN COMBINING ACRCLOUD’S NEWEST TECHNOLOGY WITH OUR MULTIPLE IN-HOUSE DETECTION SYSTEMS, WE HAVE DEVELOPED A TRUE BEST-IN-CLASS TECHNOLOGY SOLUTION, ENABLING US TO VERIFY OWNERSHIP AND IDENTIFY POTENTIAL ISSUES BEFORE DISTRIBUTION WITH THE SUPPORT OF A TRUSTED INDUSTRY LEADER.”
NICHAL SETHI, SOUNDON
Nichal Sethi, Head of SoundOn, EMEA, said: “SoundOn was built to support artists at every stage of their journey, from first release through to global success, combining distribution, promotion and audience development in one platform. As we continue to help hundreds of thousands of artists grow their fanbases and generate revenue both on and off TikTok, it’s essential that we uphold the highest standards of responsibility across the ecosystem.”
Added Sethi: “In combining ACRCloud’s newest technology with our multiple in-house detection systems, we have developed a true best-in-class technology solution, enabling us to verify ownership and identify potential issues before distribution with the support of a trusted industry leader. This ensures that the music flowing through SoundOn is original, authorized, and positioned to deliver long-term value for artists and rightsholders alike.”
The partnership comes amid heightened industry scrutiny of manipulated audio distributed through open platforms — the very type of content that ACRCloud’s Derivative Works Detection service is designed to catch.
TikTok’s decision effectively forced Merlin’s 500+ member labels and distributors to accept direct agreements or lose their ability to monetize music on the platform.
As part of the agreement, TikTok pledged protections for UMG‘s roster on the platform, including against AI-made and AI-manipulated music — underscoring the commercial pressure on TikTok, and by extension SoundOn, to demonstrate robust anti-fraud capabilities.
The issue of distributor accountability for manipulated content has since escalated further.
In November 2024, UMG, ABKCO Music & Records and Concord Music Groupfiled a $500 million copyright infringement lawsuit against Believe and its subsidiary TuneCore, alleging that Believe had built its business through “industrial-scale copyright infringement,” in part by distributing sped-up and manipulated versions of copyrighted recordings to platforms including TikTok, Spotify and Apple Music.
ACRCloud — which also provides broadcast monitoring, second screen interaction, copyright compliance, audience measurement, data de-duplication and offline recognition solution with audio fingerprinting technologies — is an existing partner of SoundOn.
Commenting on the expanded partnership, Tony Li, Co-Founder of ACRCloud, added: “We’re proud to partner with SoundOn to bring our automatic content recognition technology to a rapidly growing global distribution platform.
“By enabling accurate identification and verification of music at scale, we’re helping support a more transparent and reliable ecosystem for artists, rightsholders, and digital services alike.”
“BY ENABLING ACCURATE IDENTIFICATION AND VERIFICATION OF MUSIC AT SCALE, WE’RE HELPING SUPPORT A MORE TRANSPARENT AND RELIABLE ECOSYSTEM FOR ARTISTS, RIGHTSHOLDERS, AND DIGITAL SERVICES ALIKE.”
TONY LI, ACRCLOUD
For SoundOn, the ACRCloud partnership also comes as it continues to see growth in artist registrations since its launch in 2022.
In September, SoundOn said it had seen more than 1.1 million artists register, of whom “hundreds of thousands” have released music and generated revenue on the platform. The service expanded into Germany in September, having already been available in the US, UK, Australia, Brazil and Indonesia.
Meanwhile, the music industry continues to grapple with the influx of AI-generated content. In January 2026, Deezerdisclosed that it was receiving approximately 60,000 fully AI-generated tracks per day — around 39% of all daily deliveries to the platform.
Up to 85% of streams on that content were detected as fraudulent, demonetized, and removed from the royalty pool, according to the company.
In January, Deezer revealed that it was planning to license its AI-detection tech to the wider music industry and made it available via its revamped Deezer for Business unit earlier in March. Most recently last week, Deezer struck a deal with Hungary-based music rights organization, the Bureau for the Protection of Performers’ Rights (EJI), to license its AI detection technology.
Also last week, Musixmatch launched a service to detect when copyrighted music or lyrics are used in AI-generated and user-generated content.
The Bologna-headquartered lyrics and music data company’s new product called Sentinel is described as a lyric fingerprinting service that launched initially with the ability to detect even partial use of copyrighted lyrics “in milliseconds,” drawing on what it calls the world’s largest verified music and lyrics database. Musixmatch says the tool allows platforms to act on potential infringements in real time.
Elsewhere, Sony Music in March asked streaming platforms to take down more than 135,000 songs it says were created by fraudsters using generative AI to impersonate artists on its roster.
