Spotify is cracking down on music fraud by removing 75 million songs from its catalog, which it considers spam. This comes in response to the flood of fraudulent audio tracks submitted to the service. This is a manipulation of the royalty system or “AI garbage,” as they call it. In a new post on “For the Record,” the streaming platform outlined several measures designed to protect the music ecosystem. It also announced a list of new policies prohibiting impersonation, spamming, and tracks used to generate fraudulent music.
“At its best, AI opens up incredible new ways for artists to create music and listeners to discover”
Spotify and Deezer have teamed up to crack down on this massive fraud, which has turned out to be even bigger than many imagined. “At its best, AI opens up incredible new ways for artists to create music and listeners to discover it,” Spotify stated. “At its worst, AI can be used by malicious actors and content farms to deceive listeners, introduce spam content into the ecosystem, and interfere with the professional development of authentic artists.” Deezer estimated the number at 150,000, with 28% of them being generated solely by AI.
One of the biggest problems facing artists is AI-powered vocal cloning
In terms of specific data, music data company Luminate tracks consumption data for more than 200 million songs worldwide. By removing 75 million songs, Spotify has significantly reduced its catalog. One of the biggest problems facing artists is AI-powered vocal cloning, where fake voices are created to imitate famous singers. This is especially serious because it can also pose legal issues of impersonation, not to mention the massive music fraud this entails. Because of this, Spotify announced a new impersonation policy that only allows vocal impersonations if the original artist has given permission.
Spotify will soon implement a new AI-based spam filter
At the heart of this announcement is the immense and growing volume of music being submitted to streaming services daily. With the prevention and restrictions now in place, Spotify hopes to block scams before they reach listeners. Based on the reports, Deezer claims to have the ability to detect songs generated entirely by AI with great accuracy; it states that 70% of these songs are fraudulent (e.g., attributed to fake artists). To combat this, Spotify will soon implement a new AI-based spam filter. This system will detect suspicious users uploading content and ensure that legitimate creators are not punished for this type of impersonation.
AI can instantly generate fake metadata, such as information about artists and rights holders
AI is a complex system that can both help eliminate fraud and create it. AI can instantly generate fake metadata, such as information about artists and rights holders, and massive amounts of junk tracks, which also contribute to system manipulation and fraud. Spotify is working with Digital Data Exchange (DDEX) to create new metadata standards. These standards will help artists clearly and transparently recognize the use of AI, and the information will be displayed in the Spotify app. Digital distributors also face overhead issues, as well as potential legal liabilities, stemming from these tactics. Therefore, it is especially important to implement measures that eliminate this fraud, not least in the interests of the companies that distribute the music.
What we can deduce is that, on the one hand, digital distribution is a highly competitive market, and on the other, large companies are racking their brains trying to use AI to their advantage. The complexity of artificial intelligence is a fact, and knowing how to use it in such a way that it works for us and not the other way around is like being a pirate and striking gold.
