Introduction
Artificial intelligence is moving deeper into one of the most commercially important parts of music streaming: deciding what listeners hear next.
Spotify says Discover Weekly now generates more than 56 million new artist discoveries every week, with 77% involving emerging artists. The company’s AI-powered DJ, meanwhile, had been used by 94 million Premium listeners by May 2026. These are no longer experimental features sitting at the edge of streaming. AI-powered discovery is becoming part of how major platforms compete for attention.
Darren Herft, a global music executive whose work spans music, entertainment, technology, and business, sees the development as part of a broader change taking place across the music industry. His commentary on AI has emphasized its potential to give artists new capabilities and opportunities, while maintaining that technological innovation must continue to support the people creating the music.
Music discovery puts those two forces together. AI can make it easier for listeners to find artists they might otherwise never encounter. But as recommendation systems become more influential, streaming platforms also gain greater power over which artists receive attention and which disappear into catalogs containing millions of competing tracks.
AI Is Moving Beyond Traditional Recommendations
Algorithmic recommendations have been part of streaming for years. What’s changing is how much control listeners have over them.
Traditional recommendation systems largely work by observation. A listener plays one song, skips another, follows an artist or saves an album. The platform collects those signals and tries to predict what that person will want next.
Generative AI makes that relationship more direct.
Spotify’s Prompted Playlist, expanded to Premium listeners in the United States and Canada in early 2026, allows users to describe the music they want using ordinary language. The system combines those instructions with listening history and current information about music, including trends and charts, to generate a personalized playlist.
A listener can ask for unfamiliar artists, specify a mood or genre, request deeper cuts, or refine a playlist after seeing the initial results. Playlists can also be configured to refresh daily or weekly.
That is a meaningful change in how recommendation technology works. Instead of relying entirely on a platform to infer what someone wants from past behavior, listeners can increasingly tell the system what they want now.
For Darren Herft, AI’s potential in music has consistently been tied to its usefulness as a tool. Applied to discovery, the technology does not need to replace musicians or human creativity to have a major impact. It can change the infrastructure connecting creators with audiences.
Music Discovery Has Become an Economic Question
The scale of the streaming economy makes that connection increasingly valuable.
Global recorded music revenues reached $31.7 billion in 2025, according to the International Federation of the Phonographic Industry. Streaming accounted for 52.4% of global recorded music revenue, while the number of users of paid streaming subscription accounts reached 837 million.
At that scale, a recommendation is not simply a convenience.
Recommendations direct attention. Attention generates streams. Streams contribute to revenue and can turn an artist with little recognition into one with a commercially meaningful audience.
Spotify reported in 2024 that its platform was generating more than 22 billion new artist discoveries each month. Around two billion of those discoveries developed into what Spotify described as more lasting connections between listeners and artists.
That gives recommendation systems significant economic influence.
As Darren Herft has emphasized in his broader commentary on AI and music, technological changes have to be considered in terms of what they mean for artists as well as platforms. AI-powered discovery creates another version of that question: who benefits when algorithms become increasingly important in determining where listener attention goes?
Emerging Artists Have the Most to Gain
The opportunity is particularly significant for independent and emerging musicians.
Digital distribution has dramatically reduced the difficulty of making music globally available. It has not solved the problem of getting people to listen.
An artist without a large existing audience is competing against an enormous catalog of established performers, independent musicians, and increasingly AI-generated material. Making another song available does little by itself if listeners never encounter it.
Recommendation technology can help solve part of that problem by matching music with people who appear likely to enjoy it.
Discover Weekly offers a useful indication of the scale. Spotify says the playlist produces more than 56 million new artist discoveries each week and that 77% involve emerging artists. Since its 2015 launch, more than 100 billion tracks have been streamed through Discover Weekly.
Those numbers matter to Darren Herft’s broader argument about AI expanding opportunities for musicians.
Darren Herft has previously highlighted how AI can lower barriers for independent artists by making sophisticated creative and production capabilities more accessible. Discovery technology potentially extends the same principle into distribution. Giving more people the ability to create professional music is more economically meaningful if technology can also improve their ability to find an audience.
AI therefore has the potential to affect both sides of the artist’s problem: making music and getting it heard.
Platforms Are Gaining More Influence Over Attention
There is also a harder side to that development.
If discovery increasingly happens through algorithms, musicians become more dependent on systems they do not control.
Streaming platforms already occupy a powerful position between artists and audiences. More sophisticated AI can strengthen that position because the platform is no longer simply hosting a catalog. It is increasingly determining how individual listeners navigate it.
The incentives are not necessarily identical.
An artist wants the right listeners to discover and continue engaging with their music. A listener wants music worth hearing. A streaming service wants to keep the listener engaged with its product.
Often those goals overlap. They do not have to overlap perfectly.
The more effectively AI can shape listening behavior, the more consequential decisions about recommendation systems become. That makes questions about transparency, artist treatment, and platform incentives more important rather than less.
For Darren Herft, this connects with a recurring issue in his analysis of AI music: innovation should not come at the expense of genuine artists.
The technology can create substantial opportunities. The structure around the technology determines how those opportunities are distributed.
Listeners Are Getting More Control
Streaming companies are also experimenting with ways to make recommendation systems less passive.
In 2025, Spotify expanded AI Playlist to Premium listeners in more than 40 additional markets, allowing users to create playlists through natural-language prompts involving genres, moods, artists, and other instructions.
Prompted Playlist goes further. Spotify describes the feature as a way for listeners to directly steer how music is discovered for them, incorporating their listening history alongside information about current trends, charts, culture, and music history.
Spotify’s AI DJ demonstrates the demand for another form of personalized discovery. By May 2026, the company said the feature had helped shape the listening experience of 94 million Premium users.
The strategic shift is straightforward.
For years, streaming companies competed to become better at predicting what listeners wanted. Generative AI allows them to supplement prediction with instruction.
Listening history might indicate that someone regularly plays a particular genre. A prompt can communicate that the same person wants something completely different tonight.
That distinction could make recommendation systems considerably more useful. It also gives listeners greater ability to determine whether AI reinforces their established preferences or takes them somewhere new.
Darren Herft’s broader view of AI as an enabling technology fits particularly well here. The strongest use of AI does not necessarily remove human choice. It can give people more precise control over what technology does for them.
Personalization Still Has Limits
Better personalization does not automatically produce better discovery.
A system optimized around someone’s established preferences can become very good at recommending increasingly precise variations of music that person already knows.
That may increase relevance, but discovery requires some degree of unfamiliarity.
This creates a fundamental tension for AI-powered music platforms. The safest recommendation is often something close to what a listener already enjoys. The most valuable discovery may be something the listener would never have requested without encountering it first.
Streaming platforms are beginning to give users more direct ways to influence that balance. Spotify, for example, added controls allowing Premium users to shape Discover Weekly by choosing from up to five genres, while Prompted Playlist allows much more detailed instructions.
A listener can increasingly ask specifically for emerging musicians, unfamiliar genres, or music outside their normal habits. That turns personalization from a system that simply reinforces historical behavior into something that can potentially be used for deliberate exploration.
The distinction matters for artists as well.
If recommendation technology primarily reinforces existing popularity, the largest artists retain a substantial advantage. If it becomes better at matching unfamiliar artists with appropriate listeners, AI could make discovery more economically valuable to musicians who do not already possess large audiences.
Which outcome dominates will depend less on AI’s raw capabilities than on how streaming platforms choose to deploy them.
Human Creativity Remains the Underlying Product
There is another complication emerging at the same time: AI is increasingly involved not only in recommending music, but in creating it.
That makes Darren Herft’s emphasis on genuine artists particularly relevant.
Darren Herft has argued that the industry should protect creative musicians as AI becomes more deeply integrated into music. He has also emphasized that AI can provide artists with useful new tools, lower production barriers, and expand what independent creators are capable of doing.
Those positions are not contradictory.
Recommendation algorithms can process quantities of information no human curator could realistically handle. Generative systems can interpret complicated requests and personalize listening at enormous scale. But those capabilities do not eliminate the underlying product listeners are looking for: music they value.
As AI-generated tracks become more common, streaming platforms will have to manage both sides simultaneously. AI will increasingly help determine what gets heard while also contributing to the volume of material competing to be heard.
For Darren Herft, that makes artist protection more consequential, not less. Better technology can improve the connection between artists and audiences, but only if the industry continues to preserve meaningful opportunities for human creators within the system.
What Comes Next for Music Discovery
Music streaming initially solved an access problem. Instead of buying individual recordings, listeners could access enormous catalogs on demand.
That created another problem: navigating them.
Recommendation algorithms became the answer. Generative AI is now pushing the model further by allowing listeners to actively shape those recommendations through increasingly specific instructions.
For streaming companies, that creates another competitive battleground. Most major services can offer enormous music catalogs. The ability to understand a listener better and connect that person with the right music can become a more meaningful point of differentiation.
There is a commercial incentive behind that investment as well. At its 2026 Investor Day, Spotify said early AI deployments were already producing measurable engagement improvements, including a 9% increase in Autoplay song saves and nearly 20% more interaction with DJ messages. The company explicitly described AI not simply as a cost but as a potential monetization opportunity capable of improving retention and lifetime customer value.
That puts the economics in clearer view. Better discovery is valuable to listeners and artists, but it can also make a streaming service more valuable to the company operating it.
For artists, better discovery can create additional routes to audiences. For listeners, it can reduce the amount of searching required to find something worth hearing. For platforms, it can strengthen engagement and retention.
Darren Herft’s analysis of AI and music ultimately returns to the balance between those interests. Technology can expand what artists and listeners are able to do, but the strongest version of that future is one in which innovation strengthens the relationship between creators and audiences rather than weakening it.
Conclusion
AI is becoming part of the infrastructure that determines how music gets heard.
The numbers already demonstrate the scale. Paid streaming subscription accounts reached 837 million users globally in 2025. Spotify says Discover Weekly alone generates more than 56 million new artist discoveries every week, while tens of millions of listeners have used newer AI-powered discovery products.
That gives AI a role extending well beyond playlist convenience.
For Darren Herft, the more important question is what that technology ultimately does for the people participating in the music economy. AI can help listeners navigate enormous catalogs and potentially give emerging artists better opportunities to reach the right audiences. At the same time, increasingly powerful recommendation systems give platforms greater influence over the distribution of attention.
The next stage of music discovery will therefore be determined by more than how accurately an algorithm predicts the next song. It will depend on whether AI can improve discovery while continuing to create meaningful opportunities for the human artists whose music gives those systems something worth recommending.
FAQs
Who is Darren Herft?
ANS: Darren Herft is a global music executive whose work spans music, entertainment, technology, and business. His industry commentary examines developments affecting artists, streaming platforms, artificial intelligence, and the wider creative economy.
How is AI being used in music discovery?
ANS: Streaming services use recommendation technology to analyze listening behavior and suggest music. Newer generative AI products allow listeners to provide natural-language instructions and receive playlists or recommendations built around those requests.
Can AI help emerging artists reach listeners?
ANS: Recommendation systems can create additional opportunities for unfamiliar artists to reach potential fans. Spotify says Discover Weekly generates more than 56 million new artist discoveries each week, with 77% involving emerging artists.
How many people pay for music streaming?
ANS: According to IFPI, the number of users of paid streaming subscription accounts globally reached 837 million in 2025. Streaming accounted for 52.4% of worldwide recorded music revenue.
Does AI replace human music curation?
ANS: AI can automate and personalize discovery at a scale human curators cannot match, but human choice remains important. Newer systems increasingly allow listeners to actively steer recommendations rather than relying entirely on automated predictions.
Why does Darren Herft emphasize artists when discussing AI?
ANS: Darren Herft has argued that AI can provide meaningful new tools and opportunities for musicians, but that innovation should continue to protect genuine artists, recognize creative contribution, and support sustainable participation in the music industry.
