Spotify Removed Tens Of Thousands Of Songs To Combat Fraud — Not Artificial Intelligence

In the first week of 2021, an unsettling report went viral online: According to an entertainment attorney, Spotify removed “some 750,000 songs” from its platform due to evidence of streaming fraud, and “the vast majority” of the songs “appear[ed] to have used Distrokid for distribution.”

On social media, outraged artists blasted both the streamer and the distributor, claiming they had never been part in fraudulent streaming and had no idea why their music had been removed.

Spotify, on the other hand, denied both claims, claiming that the amount of tunes removed was significantly less than 750,000 and that music from a range of distributors was impacted. According to Distrokid’s founder, Philip Kaplan, “these takedowns were distributor-agnostic and affected music from all distributors (not just DistroKid).” Despite these rebuttals, the initial assertions persisted for months.

A variant of this occurrence occurred more than two years later. Boomy, a music-tech startup that allows users to compose songs with the assistance of artificial intelligence algorithms, announced on its Discord channel on May 1 that new uploads had been suspended and “certain catalog releases” had been removed from Spotify due to “potentially anomalous activity.”

The Boomy post was measured, but the reaction was not: The company’s remark was initially seen as confirmation that AI was causing further problems in the music industry, amid a wave of anti-AI sentiment. The story then shifted: Spotify stated that the “anomalous activity” was due to streaming fraud, not the fact that Boomy’s technologies use artificial intelligence. Then it was revealed that Spotify had removed more songs — which had nothing to do with Boomy or AI — due to signs of manipulation discovered during a normal scan a few days later.

The entire problem now appears to be less about any one firm and more about the continuous efforts of streaming services to prevent fraud from affecting payouts on their platforms. (Over the years, Spotify has regularly stated that “stream manipulation is an industry-wide issue that” it takes “very seriously.”)

These episodes demonstrate the difficulty of accurately reporting on the murky world of streaming fraud, where even the most basic information — how many tracks were impacted, what criteria were used to determine whether they were manipulated, and how it compares to overall fraud levels — is frequently kept out of reach by tech companies. However, combating streaming fraud is a never-ending game of whac-a-mole that affects all streamers and distributors to varied degrees, and focusing on any particular mole might obscure the greater picture. As Christine Barnum, chief revenue officer at CD Baby, recently told Billboard, “nobody’s immune” to this type of deception.

Boomy claims that around 7% of its music on Spotify was removed because those songs were targeted with bot activity in April; following a brief halt, Boomy users were able to continue uploading new songs to Spotify on May 5. Last year, a Deezer executive stated that “7% of the volume of daily streams [on Deezer] is now detected as fraudulent.” In 2020, Merlin, which manages digital licensing for many famous independent labels and distributors, experienced a brief spike in fraud levels reaching 10% (from music on Spotify’s ad-supported tier).

Speaking publicly about streaming fraud was once considered “airing dirty secrets,” one executive recently told Billboard. However, this is changing: SoundCloud, Pandora, and Napster executives all spoke on a Music Biz conference panel in 2022 about their efforts to combat fraud on their respective platforms. Last month, Umeadi Onyekwelu, music licensing head at Mdundo, said that “looking at the music industry in Nigeria, one of the biggest problems is stream farming, which has become more widespread and prominent over the years.”

This is the modern music ecosystem’s day-to-day reality. Deezer claimed last year that it detects unusual behavior on 45,000 accounts per day, while Spotify sends frequent reports to key rights holders about the extent of fraud identified on their catalogs. (Because the fraudulent plays revealed in those reports were caught, they had no effect on rewards.) According to a Spotify spokeswoman, the site “consistently removes products designed to game the system in order to generate royalties.”

Boomy’s CEO, Alex Mitchell, told Billboard this week that “our review team spends a huge amount of time on that issue [protecting the platform from fraud].” To be honest, we have considerably harsher regulations than many other distributors. We have procedures in place that inform us if we suspect something is odd. Then our team will conduct an investigation to determine whether they need to hold the revenue or contact a DSP. We’re also collaborating with industry-leading fraud detection firms to improve our systems as we grow.”

Nonetheless, anyone’s music on a streaming network can be targeted by bots. Morgan Hayduk, co-founder/co-CEO of Beatdapp, a firm that develops software to detect and reduce fraud, told Billboard earlier this year that “collateral damage” was an under-discussed component of fraud. “Fraudsters frequently employ user accounts on the streaming platform to stream a mix of the target’s content alongside other popular artists to evade detection,” he stated, referring to bots that play a mix of music from real musicians, as well as the track they are attempting to enhance. (Beatdapp collaborates with Boomy but refers any “anomalous activity” to Boomy.)

This implies that, even when fraud is detected, determining its origins might be difficult. “It’s difficult to pinpoint who is responsible,” Ludovic Pouilly, senior vp of institutional and music industry partnerships at Deezer, told Billboard in January. “Distributors may claim that it’s the labels. It could be the management, according to the labels. And the artists themselves may claim that the competition is attempting to harm their reputation.”

At a time when more platforms are openly debating fraud, the Boomy announcement on Discord undoubtedly drew a lot of attention due to the company’s ties to artificial intelligence, a topic that looks to be making many music executives nervous. (“People are over-panicking a little bit,” said one seasoned music tech executive to Billboard recently.) Some of that fear stems from the dilution of major label market share, which affects payouts from streaming services. The music business is also concerned about the possibility for copyright infringement and the extent to which AI technology could potentially replace musicians and songwriters.

However, as Mitchell points out, streaming fraud “existed for a long time before AI was on the scene, and before we were on the scene.” Fighting fraud was already a tech-based arms race between those who want to safeguard the royalties pools of streaming services and those who want to take money from them.

According to a January French study on streaming fraud, “the imagination of hackers is rich and evolving, to the point where the countermeasures imagined and implemented by platforms in the first place — but also distributors and music rights holders — must not only constantly evolve and improve, but also anticipate any counter-offensive from fraudsters.”

Boomy may have summed up the situation best when it first announced that Spotify had discovered potential evidence of fraud on several releases. “As the music industry navigates the use of bots and other potentially suspicious activity,” the company noted, “these pauses are likely to happen more frequently and across a broader set of platforms.”

Spotify Removed Tens Of Thousands Of Songs To Combat Fraud — Not Artificial Intelligence

Download The Radiant App To Start Watching!