Spotify has conducted a massive cleanup of its platform, removing approximately 75 million compositions generated by artificial intelligence. This move is aimed at curbing a growing wave of fake, mass-produced songs uploaded solely to manipulate the royalty system and divert income from genuine artists.
The company emphasized that the issue is not with the use of artificial intelligence itself, but with its misuse. According to Spotify's Senior Director, Sam Duboff, the platform is fighting what the industry calls 'AI slop'—low-quality tracks generated in seconds and uploaded in huge volumes with minimal creative input.
Modern AI tools allow for the creation of complete songs from a simple text prompt, enabling malicious actors to flood streaming services with thousands of nearly identical compositions. This flow is growing rapidly: about 100,000 new songs are uploaded to Spotify daily, and industry estimates suggest that almost half of all new music on streaming services is now created with AI.
The motivation of fraudsters is simple—financial gain. Since every listen contributes to the overall royalty pool, violators exploit this system by uploading gigantic libraries of generic ambient sounds, meditative music, and repetitive instrumental pieces. They then use automated bots to generate artificial streams.
These songs are often published under fictitious artist names and are intended for background playlists where listeners are unlikely to notice who is performing. As a result, money intended for authentic musicians is diverted to those who have distorted the system. For independent artists from South Africa, where every stream holds significant weight, such fraud carries real financial consequences, as every fake stream reduces the share available to musicians who invested time, talent, and resources into creating original music.
Spotify has clearly stated that AI-generated music is not automatically prohibited. The company continues to permit the use of AI-generated music provided that the creators hold the appropriate commercial rights and the technology is part of a legitimate creative process. The focus remains on abuses.
The platform has enhanced its machine learning systems to detect suspicious upload patterns and streaming activity related to bots. Furthermore, unauthorized voice cloning using AI has been banned, meaning tracks imitating the voices of famous artists without permission may be removed. Spotify has also implemented additional safeguards to prevent the unauthorized scraping of the company's catalog for training AI models.
The discussion about artificial intelligence in music is far from over. Music industry organizations, including IFPI and RIAA, are pushing for clearer labeling to indicate whether songs are AI-generated or AI-assisted. Proponents argue that greater transparency will help listeners make informed choices and protect the work of human artists.
As AI technologies advance, streaming services will face increasing pressure to distinguish between genuine creativity and content created solely for algorithmic exploitation. The removal of 75 million tracks by Spotify was an important step, but likely not the last. The challenge is to stay ahead of those who intend to cheat the system. For music lovers, this cleanup serves as a reminder that behind every legitimate stream there is a songwriter, producer, or performer whose livelihood depends on fair compensation. Protecting these creators may prove as vital as adopting the next generation of music technology.
Spotify has begun testing Talk to Spotify, an artificial intelligence-based feature that allows users to create playlists, ask questions about songs, and access their own listening history data using both voice and text.
The company announced that this new feature will first be available to Premium subscribers. The central proposal is to simplify interaction, allowing the user to converse directly with the platform instead of navigating through various internal functions of the application. The functionality simulates a conversation, where the subscriber can request suggestions, modify music selections, and obtain more details about artists without leaving the application.
According to Spotify, it will be possible to maintain a continuous dialogue to define the content to be played, deepen knowledge about music, and explore podcasts and audiobooks. Talk to Spotify is accessible in the Home and Now Playing sections of the mobile apps. To start the interaction, simply press the microphone icon in search or type a command. Among the supported actions are creating personalized playlists, changing the style of a selection, saving tracks, adding songs to the queue, and tracking singers.
In addition to assisting with track curation, the AI acts as a source of musical information. Through the playback screen, the user can inquire about a specific song, album, or artist. For example, it is possible to ask about the inspiration behind Dua Lipa's album 'Radical Optimism', or identify the genre of a particular track at the time of its release. This conversation can also generate new discoveries, suggesting related artists and providing data on podcasts and audiobooks. Furthermore, the user has access to details of their own journey on the platform, such as the date of first contact with a song or the most recently listened-to styles.
Talk to Spotify is being distributed in beta version to Premium subscribers aged 18 or older in the United States, Ireland, and Sweden. The test operates on Android and iOS devices, starting with English support. Spotify already uses artificial intelligence in other features, such as remixing and generating podcasts on command. With this new test, the company introduces a more conversational way of engaging with the service. Spotify clarified that the function allows for a direct exchange with the platform, enabling the user to learn about their tastes, review their listening history, and dive into podcasts and audiobooks, all within Spotify itself.
Spotify registered stability issues during the morning of this Tuesday, July 14. According to information released on social media, both the desktop applications and the browser version experienced slowness when loading the service's main page. At previous times, the loading was incorrect.
The occurrence of the disturbances was detected around 8:30 AM, as recorded in the DownDetector website chart, a platform specialized in monitoring the performance of online services. Complaints reached their peak around 9:00 AM but began to decrease rapidly, returning to near-normal levels by 9:30 AM.
Reports made on the X platform (formerly known as Twitter) confirm that the failure was not restricted only to Brazil, with mentions in Portuguese and several other languages.
