In late 2025, Papaoutai, a French-language song, exploded across streaming platforms. The song currently has close to 140 million streams on Spotify.
What most listeners did not realise was that the viral afro-soul version remix of the 2013 hit by Belgian artist Stromae was created using AI, and with most music platforms lacking clear labelling for AI music, many listeners did not know.
AI-generated tracks now account for 44% of all new music uploaded to the platform, according to streaming platform Deezer, and 97% of people it surveyed could not hear any difference between AI- and human-made music.
As AI-generated music becomes harder to distinguish from human-created work, streaming platforms are being forced to answer uncomfortable questions about impersonation, fraud, ownership, and authenticity.
For Spotify, this means introducing new safeguards to promote AI transparency on its platform.
“We believe that to truly unlock the positives and the potential in AI, we need to protect against the worst,” Bryan Johnson, Spotify’s Head of Artist & Industry Partnerships, told TechCabal during a two-day event at the company’s new South African office in Johannesburg on May 14.
One of Spotify’s biggest concerns is spam. Johnson said the company removed 75 million spammy tracks from the platform in the past year as AI tools made it easier to mass-produce low-quality or deceptive uploads.
In September 2025, Spotify explained that its spam filter was created to identify uploaders and tracks engaging in spam tactics, including mass uploads, duplicates, SEO hacks, artificially short track abuse, and other forms of slop. The system was designed to tag these tracks and stop recommending them.
The company is also overhauling its artist verification process to what it describes as a more human-centric system.
Previously, artists who signed up for Spotify for Artists automatically received a verification badge. Under the new Verified by Spotify system, artists must now demonstrate sustained listening activity, avoid fraudulent behaviour, and show evidence of real-world artistic activity such as ticket sales, performances, or merchandise sales before receiving verification.
“We want regular listening, so sustained listening activity, which I think is at least 10,000 monthly active listeners over three consecutive months,” Johnson said during a panel session.
Spotify is also testing what it calls AI credits, a feature that would allow artists and labels to disclose whether AI was used during parts of a song’s creation, including songwriting, instrumentation, or production. The platform is currently working with a few distributors on this.
“The information is delivered from the artist or songwriter to the distributor label, and they deliver it to Spotify, and we can surface that on the artist page,” Johnson said. “It is all about giving artists more control over their profile, their presence on the platform in this AI era, and giving listeners more trust.”
For now, Spotify insists AI is not a problem.
The company has rolled out several AI-powered listener features, including AI DJ, a personalised, AI-powered music guide, which Johnson said has reached about 94 million listeners since launch.
“We are leveraging this technology to give the best experience for listeners and to give them more control over the platform,” he said. “On the artist side, we are taking protective measures there too.”
Spotify’s changes reflect a wider industry shift. Deezer began tagging AI-generated tracks in June 2025 and says it tagged more than 13.4 million AI tracks on its platform that year alone.
As synthetic voices become increasingly indistinguishable from human ones, streaming companies are moving beyond simply hosting music and are beginning to decide what authenticity should look like in the AI era.