Spotify, with its massive library of over 100 million tracks and more than 600 million users, faces a challenge: how to help listeners discover music they’ll love. Their goal is personalization and meaningful recommendations, making their vast catalog more valuable to users.
To achieve this, Spotify has developed various recommendation tools like the Home feed, Discover Weekly, Blend, Daylist, and Made for You Mixes. These efforts have shown success, with artist discoveries increasing from 10 billion to 22 billion monthly by 2022.
Investing heavily in AI and machine learning, Spotify recently launched its AI DJ, a tool aimed at enhancing personalization and introducing users to new music. This AI DJ mimics radio vibes, announcing songs and introducing tracks to help users step out of their comfort zones.
Behind the scenes, Spotify combines personalization technology, generative AI, and dynamic AI voices. Human experts, including music editors and tech specialists, work to improve recommendation tools using generative AI to scale their expertise.
Data on songs includes musical features, release year, genre, and mood, allowing Spotify to generate new recommendations by analyzing millions of listening sessions. The AI works on the principle of “Users who liked Y also liked Z,” effectively matching users with similar tastes.
According to Spotify, the AI DJ has encouraged listeners to explore new music by providing commentary alongside recommendations. This benefits not only users but also artists seeking to connect with new fans.
Julie Knibbe, CEO of Music Tomorrow, emphasizes the challenge of balancing familiarity and novelty in music discovery. While AI can predict preferences well, it struggles to understand when users want to explore new genres or styles.
Spotify’s Daylist attempts to address this by using generative AI to recommend music based on users’ varying moods and activities throughout the day. However, Knibbe acknowledges that not everyone wants to discover new music constantly; many users prefer familiar listening patterns.
Ben Ratliff, a music critic, believes AI algorithms can simplify music taste and perpetuate clichés rather than enhancing discovery. He prefers curated playlists made by individuals with genuine preferences.
Ultimately, whether AI enhances or detracts from the music discovery experience depends on the user. Ratliff advises users to keep their music journeys simple and understand that no app can fully know their preferences.
Source: CNBC
