SoundCloud’s Attempt at a Revamp? An AI-Driven Recommendation Feed

AI has infiltrated just about every creative field. Poets have complained about the tech’s shoddy imitations of famous writers, anime fans can watch an unending AI-generated show and artists are suing an AI company over copyright usage. The music industry is no exception.

Though there are plenty of examples of people using ChatGPT to write songs, AI has been most successfully implemented as a way for music platforms to recommend music.

Yesterday, SoundCloud announced an AI-driven, TikTok-esque recommendation feed. Based on tech from Musiio, which SoundCloud acquired in 2022, the feed shares 30-second clips of songs inspired by users’ listening history and music taste. This comes not long after Spotify launched its own vertical AI feed that users can scroll through to discover new music. In addition, Spotify has gotten into the AI game with its new DJ—a personalized music feed accompanied by an AI-generated voice reading facts curated by Spotify’s experts.

For SoundCloud, this move is likely an attempt to revamp a struggling business. Last August, SoundCloud cited "the challenging economic and financial environment" when it laid off 20% of its workforce. Its business model has been questionable for quite some time, with reports going back to 2014 noting that its large user base doesn’t offset the cost of running the platform. However, the company has seen a steady revenue increase since 2017. Still, apps like Spotify and TikTok have cracked the code of music recommendations. SoundCloud’s TikTok imitation, complete with 30-second song clips, shows that the company is vying for attention.

But the message coming from these companies is a fairly obvious one: users can’t be bothered to choose the music they want to listen to for themselves.

On the surface, the release of these recommendation features make sense. We’re currently seeing more music being released than ever before. In 2022, 100,000 songs were uploaded across the various digital service providers every day. People are clearly looking for ways to parse through the wide offering. Spotify’s Discover Weekly, which suggests new music to users, was streamed for over 2.3 billion hours between 2015 and 2022.

While AI can be helpful for some listeners, it doesn’t bode well for music creation. Just look at what’s happened to the music industry as artists chase TikTok’s algorithm. Music executives have admitted to encouraging producers to shorten songs in order to better prime them to go viral on TikTok. This strategy has worked for some artists, like Gayle, who turned TikTok vitality into a Grammy nomination. But artists have been lambasted for valuing TikTok virality over quality. When Meghan Trainor released a preview of her upcoming single “Mother” last week, people were quick to label it “TikTok vomit.”

But it's one thing for a social media platform to boost a certain type of short, quippy song. It’s another thing entirely for the very platforms that these songs live on to do so. Even artists who currently look down on this songwriting style might be encouraged to produce similar music if it’s the only way to reach new audiences.

Every tech company is chasing the AI craze, so it's no surprise that SoundCloud and Spotify are doing so as well. The recent announcement from Soundcloud also serves as an indictment of how the attention economy has upended the music industry. One researcher found that people only pay attention to one screen for about 47 seconds. That presents a problem for music platforms when pop songs, for example, average around three minutes and 30 seconds. Hence Soundcloud’s decision to serve up songs in 30-second sound bites.