The Art of Taste-Building with UX and AI

In an age where algorithms increasingly control what we see, hear, and engage with, we’re reflecting on one big question: how do we balance personalization with the human desire for exploration and connection? Our reflection goes beyond convenience. We believe the world would be more beautiful if we craft meaningful experiences that empower users to grow and discover. Particularly in music, an art that lends itself to taste-makers and social recommendations, we’ve noticed that we might have steered too far away from making clear moments of human connection.

This post was inspired by our recent conversation on the Pragmatics Podcast about how modern music discovery might be obscuring human connection, despite the fact that it can suggest great content. Watch the video below or find it wherever you listen to your podcasts.

When Machines Curate: How AI Is Redefining Musical Taste

The Double-Edged Sword of AI Personalization

AI’s ability to analyze staggering large datasets has unlocked so many possibilities for tailoring experiences. In platforms like Spotify, algorithms monitor every track someone listens to and define our taste profiles to deliver hyper-personalized tracks, albums, artists, and playlists. Done well, this can lead to magical moments of discovery, like in the incredible recommendations in Spotify’s Discover Weekly and Daylist playlists. Yes, it delights users with eerily accurate recommendations, and yes, reduces the fatigue of landing on terrible music, but there’s a flipside. These systems can risk narrowing our horizons, locking us into echo chambers of familiarity.

For UX designers, this gives us a juicy challenge: how can we modify these systems to encourage more organic exploration before we go too far in the direction of relying entirely on machines? You might be wondering why that’s even important. What’s the problem with 100% machine curated music? The real answer is that there isn’t inherently anything wrong with it. Perhaps some people would love hyper-personalization. Yet this option’s worrying side-effects could be that people miss out on opportunities to change and grow.

Here’s an easy way to realize how much we change our tastes and grow out of certain preferences: Go back to the music you listened to during high school. Did that Blink-182 or Green Day make you cringe? I could barely finish typing this sentence thinking about my emo phase, and I am grateful everyday that I had the opportunity to be exposed to things totally outside of my comfort zone. Doing this won’t be easy. It will require giving listeners tools to finetune the balance of control and curiosity.

Designing for Discovery

The history of discovering music has always been rooted in community. We relied on word of mouth, mixtapes, serendipitous concerts with friends, and forums to uncover new ideas and expand our tastes. In contrast, today’s algorithm-driven platforms often operate in isolation. They prioritize individual engagement metrics over collective experiences. Yes, we are aware of how Spotify’s Discover Weekly relies very heavily on human intervention, but those people are invisible to the listener. Even finding user profiles on Spotify is a challenge.

This is where UX can play a big role by reintroducing elements of community into the digital space. With some clever collaborations with AI, we can reimagine systems that use AI to facilitate not just personal recommendations but shared discovery—tools that help users explore new interests collaboratively. We’re already seeing features like Spotify’s Blend playlists, but there’s room to dream bigger. In the past, full platforms like TheSixtyOne and HypeMachine were built around the idea of communities floating music from obscure depths to the mainstream. Those platforms gave listeners the ability to intentionally curate large-scale tastes rather than passively by monitoring what’s popular.

Beyond Patterns: AI as a Reflective Partner

AI’s best skill currently is pattern recognition, but that means its outputs are only as good as the data it’s trained on. This can lead to quirks and biases in how preferences are identified that feel obscure and at times even off-putting. Though it might take more effort on the users part, we see room to design AI systems that not only surface recommendations but also invite reflection.

For example, why don’t music platforms ask listeners to not only rate how a playlist aligns with their mood or long-term taste-building goals, but also give them tools to explicitly block out certain songs and vibes? This feedback loop would refine the algorithm while giving listeners the opportunity to think critically about their preferences, fostering a sense of agency over their own discovery journey.

Would these kinds of user interfaces mean more effort on the users’ part? Perhaps! But we know that users wanting more control is not unique. In a previous talk I’ve given, I highlighted how Netflix replaced the 5-star system with a Thumbs Up and Down combined with human behaviour, only to add a third option again because viewers felt that only 2 options wasn’t granular enough control. This is proof that our narrative around humans being lazy and not wanting to think about what they consume might be shaped more by our current business leaders more so than any underlying truth about human behaviour.

Taste-Building as a Design Imperative

Personalization should not mean pigeonholing. Perhaps there’s a way to consider the dynamic nature of human taste—our ability to grow, evolve, and seek new horizons. As designers and technologists, we should consider adding the responsibility to ensure that the systems we build don’t just predict but also inspire.

AI emerged as another tool to improve our ability to discover but, like many technologies that came before, it should not be a replacement for human insight. By combining its efficiency with reflective, exploratory interfaces, we can deliver experiences that feel deeply personal yet open-ended. Our software should empower users to shape their tastes rather than passively accept them.

The Future of AI and UX

As AI continues to redefine the boundaries of personalization, our appreciation for its capabilities will lie in how it enabled us to foster both individuality and connection. The challenge is not just technical but philosophical: how do we design systems that honor the complexity of human taste and curiosity?

This is where UX has a critical role to play. It’s not just about the data or the algorithms but about the thoughtful interfaces that frame them. In this new era, designers must advocate for tools that are not only efficient but also expansive. The systems of the future should balance the best of AI with the boundless potential of human curiosity.