Spotify on Thursday announced a series of updates to its AI policy, which was designed to indicate better when AI is being used to make music, cuts on spam, and to clarify that unauthorized voice clones are not allowed on its service.
The company says that it will adopt an upcoming industry standard to identify and label AI music in credit DDEXAnd will soon roll a new music spam filter to catch more bad actors.
Under the DDEX system, the label, distributor and music partner present the standardized AI revelations in the music credit. This solution provides detailed information about the use of AI-as it was used for an AI-generated tone, instrumentation or post-production.
“We know that AI is going to be a spectrum, AI with artists and producers included in various parts of their creative workflow,” Sam Daboff said, Spotiff’s global marketing and policy global head said in a press briefing on Wednesday. He said, “This industry standard will allow for more accurate, nuances. It will not force the track in a false binary where a song should not be either clearly AI or AI,” he said.
As part of the same declaration, Spotify clarified its polis around the AI-Enabled personalization, which directly stated that AI voice clone, deepfack, and no other form is allowed to have vocal replicas or repatriation in any other form and will be removed from the stage.
While DDEX standard is developing, Spotify says it has received 15 labels and commitments from distributors that plan to adopt technology, and see its move, it may indicate others that it is time to adopt technology.
Because AI tools make it easy to release music to anyone, Spotify also has a new plan to cut potential spams. This fall, the company will roll a new music spam filter that will try to address, tag the spam strategy, and then stop users recommending those tracks.
Techcrunch event
San francisco
,
27-29 October, 2025
“We know that AI has made it easier than ever to use SEO tricks to upload large scale for bad actors, make duplicate, search or manipulate recommended systems … We have been fighting this type of strategy over years,” Daboff said. “But AI is sharpening these issues with more sophistication, and we know that new types of mitigation is required.”
The company said it would gradually roll the filter to ensure that it is targeting the correct signals, then add more signs over time as the market develops.
Relative to this, Spotify will also work to address something called “profile mismatch” with distributors, a scheme where someone frauds uploads music to another artist’s profile in streaming services. The company said that more of these are expected to stop before the music is live.
Despite the changes, Spotify authorities emphasized that they still support the use of AI, provided it is used in a non-washed manner. Spotify VP and Global Head of Music, Charlie Helman said, “We are not punishing artists here to use AI with authentic and responsibly.” But we are here to stop the bad actors who are gaming the system, and we can only benefit from that good side of AI if we protect the negative side, “they said.
Spotify updates follow the rapid growth of AI-generated music throughout the industry. This summer, a AI-Janit Band called Velvet Sundown, went viral Leading on its service Users to complain The company is not transparent about labeling its AI tracks. Meanwhile, streaming rival Damageders recently shared that about 18% of music uploaded to its service each day-or more than 20,000 track-now is fully AI-Janet.
Spotify will not share its own metrics directly on the case – but Daboff told reporters that “reality is this, all streaming services have almost the same catalogs.”
“People distribute music to all services,” they explained, adding the uploaded tracks do not mean that someone listens or AI music makes money. “We know that AI use is not fast a binary, but how artists and manufacturers are using, it has a spectrum.”
