Problems about AI has been one of the most pressing issues of the music industry in recent months, and Spotify has announced a reinforcement of the platform’s protection against “AI slop”, spamming and other fraudulent activity. The company notes that over the past 12 months, more than 75 million “Spammy” tracks have been removed from its platform.
Its new protection includes a policy to police unauthorized vocal imitation (“deepfakes”) and fraudulent music that has been uploaded to the artists’ official profiles; An improved spam filter to prevent mass uploads, duplicates, SEO chops, artificial short tracks designed to fraudulently increase streaming numbers and payments. The company also says that it collaborates with industry partners to design an industry standard in a song’s credits to “clearly indicate where and how AI played a role in the creation of a track.”
“The pace of recent progress in generative AI technology has felt fast and sometimes worrying, especially for creativity,” writes the company in a just published post on its official blog. “At its best, AI locks incredible new ways for artists to create music and for listeners to discover it. The worst can be used by bad actors and content farms to confuse or deceive listeners, drive ‘slop’ into the ecosystem and disturb authentic artists who work to build their careers.
“The future of the music industry is being written, and we believe that aggressively protects against the worst parts of Gen AI is crucial to enable its potential for artists and producers.”
In a press meeting on Wednesday, Spotify VP and the global head of the music product Charlie Hellman said: “I want to be clear about one thing: We are not here to punish artists to use AI authentic and responsibly. We hope they will enable them to be more creative than ever.
“But we are Here to stop the bad actors who play the system. And we can only benefit from all the good side if we aggressively protect against the bad side. ”
According to the announcement, these include protection:
Stronger imitation rules
The question: We have always had a policy against misleading content. But AI tools have made it generate vocal depths of your favorite artists easier than ever before.
What we announce: We have introduced a new imitation policy that clarifies how we handle claims about AI voting clones (and other forms of unauthorized vocal impress), which gives artists stronger protection and clearer opportunity. Vocal impersonation is only allowed in music on Spotify when the mimicked artist has approved use.
We are also increasing our investments to protect against another imitation tactics-where the uploader fraudulently delivers music (AI-generated or otherwise) to another artist’s profile over streaming services. We test new preventive tactics with leading artist distributors to equip them to better stop these attacks at the source. In the end, we will also invest more resources in our content matching process, reduce the waiting time for review and enable artists to report “maladministration” even in the pre-release state.
Why it is important: unauthorized use of AI to clone an artist’s voice exploits their identity, undermines their artistry and threatens the basic integrity of their work. Some artists may choose to license their vote for AI projects – and that is their choice to make. Our job is to do what we can to ensure that the choice stays in their hands.
Music spam filter
Question: Total music payments on Spotify have grown from $ 1B 2014 to $ 10B 2024. But large payments attract bad players. SPAM tactics, such as mass uploads, duplicates, SEO chop, artificial short tracks and other forms of slop, have become easier to use because AI tools make it easier for everyone to generate large amounts of music.
What we announce: This fall, we roll out a new music spam filter – a system that will identify uploaders and tracks that participate in these tactics, tag them and stop recommending them. We want to be careful to make sure that we do not punish the wrong chargers, so we will roll out the system conservative in the coming months and continue to add new signals to the system when new schedules emerge.
Why it is important: left uncontrolled, these behaviors can dilute the royal pool and influence attention for artists who play according to the rules. Our new music spam filter will protect against this behavior and help prevent spammers from generating royalties that could otherwise be distributed to professional artists and songwriters.
AI information for music with industry standard credits
The question: Many listeners want more information about what they listen to and the role of AI technology in the music they flow. And for artists who responsibly use AI tools in their creation process, there is no way on streaming services for them to share and how they use AI. We know that the use of AI tools is increasingly a spectrum, not a binary, where artists and producers can choose to use AI to help with certain parts of their productions and not others. The industry needs a nuanced strategy for AI transparation, not forced to classify any song that either “is AI” or “not AI”.
What we announce: We help to develop and will support the new industry standard for AI information within music loans, developed through Dex. When this information is submitted via labels, distributors and music partners, we start showing it all over the app. This standard gives artists and rights a way to clearly indicate where and how AI played a role in the creation of a track-if it is AI-generated song, instrumentation or postal production. This change is about strengthening confidence in the entire platform. It is not about punishing artists who use AI responsibly or down ranked tracks to reveal information about how they were done.
This is an effort that will require broad industry orientation, and we are proud to work with this standard along with a wide range of industry partners, including: Amuse, Audiosalad, Believe, CD Baby, Discid, Downtown Artist & Label Services, Empire, Encoding Management Service – EMS GMBH, EMS GMBH Nuemeta, Nuemeta, Nuemeta, Nuemeta, Nuemeta, Nuemeta, Nuemeta, Revelator, Revelator, Revelator, Revel, Revel, Revel, Sonosu, Sourite.
Why this is important: by supporting an industry standard and helping to pursue their broad assumption, we can make sure that listeners see the same information, no matter what service they listen to. And in the end, it preserves confidence over the entire music ecosystem, as listeners can understand what is behind the music they are streaming. We see this as an important first step, which will undoubtedly continue to develop.
Posten concludes: “We support the artists’ freedom to use AI creatively, while fighting their abuse of content farms and bad actors. Spotify does not create or own music; this is a platform for licensed music where royalties are paid based on listener engagement, and all music is treated equally, regardless of tools used to do.
“These updates are the latest in a series of changes we make to support a more reliable music ecosystem for artists, for rights holders and for listeners. We will keep them as the technology develops, so stay updated.”
Amount will have more about the situation as it continues to develop.
