Making Waves in Uncharted Waters: AI Music’s Brave New World

The sound of waves crashing mingles with Ola’s voice as we sit on a Costa Rican beach, escaping the Canadian winter for a few months of remote work under palm trees. It might seem like an odd setting for a conversation about artificial intelligence and the future of music, but Ola’s relationship with AI is complicated.

“Pop music is like a religion,” she tells me, her voice carrying conviction as a frigate bird circles overhead. “We’ve all been thoroughly brainwashed by it and continue to be brainwashed by it. I think it’s the best opportunity to undo the bad brainwashing and redo it with really helpful, wholesome things.”

It’s a bold statement from someone who describes herself as “really freaked out” by artificial intelligence and thinks it’s “moving too fast.” Yet here she is, spending precious spare time when she's not working at her day-job as a UX designer, crafting AI-generated songs that emerge from a place of personal healing, a genuine curiosity about what these new tools can create, and an aspirational goal of creating music that might help humanity “do well together as a species” instead of “killing each other and destroying the planet.”

Ola embodies the contradictions many of us feel about AI right now. She’s simultaneously amazed and terrified, finding creative liberation through tools she fundamentally distrusts. Her story offers a window into what happens when ideology meets reality in an era where machines can mimic creative output, and it reveals just how complicated the act of making music with AI can be.

A Personal Path to Music

Music has been Ola’s companion for much of her life: her father was in a rock band when she was a child and built a recording studio in their basement; during her teenage years she was a club kid obsessed with alternative and electronic music; while studying fine arts in university she worked at the city’s biggest music store. Later, she channeled her passion into visual installations and live performances at underground music events, always drawn to the intersection of sound and experience.

But despite this deep musical background, she never played an instrument well and considered herself an average singer. “I had a lot of ideas,” she explains. “I even wrote a lot of little songs, and have lots and lots of little ditties in all sorts of nooks and crannies.” But translating those ideas into finished pieces wasn’t possible until AI gave her a completely new kind of instrument.

Some of her songs emerge from insights about life, and working through emotions or life experiences that benefit from musical expression. “I was actually making music as personal therapy,” she tells me, her voice carrying both warmth and clarity. “And I might as well share it with the world.”

Other songs spring from pure curiosity and joy, like when she wanted to learn Spanish verbs in the past tense and ended up creating “Vivir y Aprender,” a 90s hip-hop track that exceeded her expectations. Even her musician father, sometimes skeptical of her musical tastes, “absolutely loves it and wants to do his own version of it.”

Swimming Against a Digital Tide

This personal approach to AI music creation stands in stark contrast to what’s currently flooding streaming platforms. Deezer now receives over 20,000 fully AI-generated tracks every day—that’s 18% of all new music uploads, and nearly double the 10% reported just four months earlier. “It’s overwhelming,” Ola admits, reflecting on the numbers. “I’m contributing to this flood, but I’m trying to make sure what I add actually means something, has good vibes, and sounds great.”

The scale and sometimes deceptive nature of AI music generation has prompted dozens of musicians, including Billie Eilish, Nicki Minaj and Stevie Wonder, to warn that AI-generated music trained using their work could “sabotage creativity” and sideline human artists. Their concerns seem validated by cases like The Velvet Sundown, a band that gained over 470,000 monthly Spotify listeners while being deceptive about its use of AI, not that all of those listeners necessarily cared. Music manufactured for popularity is hardly new after all.

So it's fitting that "Blueberry Fields"—her first published EP—is a tribute to the age-old practice of remixing and riffing off of popular music, while also alluding to the surreal time we're living in. It's a collection of six songs that all contain the same lyrics, yet each one is produced in a completely different genre. Even the album art is self-referrential, incorporating a tongue-in-cheek misspelling and faux pixellation in the title.

The Reality of AI Collaboration

For those curious about the technical side, Ola’s process reveals that AI music creation involves far more than pushing a button and getting a hit song. While she may have a clear idea in her head when she sits down to create something new, the technology requires being open to flexibility and serendipity in the process.

Her songs often start with an insight from real life: “The insights happen away from computers and they’re related to real life things,” she tells me. She’ll then dig into that concept, sometimes collaborating with ChatGPT to help develop the lyrics, asking it to “make this into a rap song” or “give me ten versions of this word that rhyme, but have a more encouraging vibe.” There are many iterations and edits before it’s ready to start working with.

Then comes the technical work. Using Udio, an AI music generation platform, she’ll take her lyrics and start the cycle of “generating and assessing, generating and assessing.” She might need to tweak her prompts, change her lyrics and instructions, edit genre tags, or try negative prompting to exclude unwanted elements like horns or vocals—though she notes these exclusions “don’t always work. Sometimes it does the opposite.”

The most maddening part comes during the assembly process. Since Udio generates music in either 30-second or 2-minute increments, creating a full song means stitching together pieces while listening to the same snippets “over and over and over and over.” And once a song is structurally complete there’s often inpainting (i.e., regenerating) parts of the song that need tweaking or changing, along with cropping, extending, or shortening… it can be a fiddly and drawn out process.

A single song might take months to complete, as she often needs to step away, get a break from it, and focus on other commitments. There’s also the album artwork to consider, song titles to obsess over, and final mastering to ensure the sound levels are balanced. Then there’s putting it up on the platforms, and ideally sharing it on social media and with friends and family, though she doesn’t do that as often as she thinks she should. Anyone who thinks publishing music is easy has no idea how much thought and care goes into each release, at least for artists who care to make something worthwhile and original.

The Social Media Dilemma

Beyond any potential legal concerns associated with AI-generated music—Udio and Suno are currently being sued by the major labels for copyright infringement—Ola worries about public reception. “There’s this vibe of hating on AI that’s going on right now,” she notes, while considering how to share her music without inviting controversy.

Some platforms now require creators to declare if content is AI-generated, partly due to concerns about misinformation and deepfakes. It also prevents her music from being monetized on certain platforms like Instagram and TikTok, despite those platforms being rife with AI-generated content.

For now, she’s decided to keep a low profile. Sharing her music with close friends and family, and engaging in some light promotion on Bluesky, allows her music a chance to spread organically, if slowly. Eventually, the hope is, people will come to appreciate her songs for their vibe and their message, and it won’t matter what tools she used to create them.

Finding Harmony in the Chaos

Ola’s story reveals AI music creation as neither the effortless magic promised by tech evangelists nor the creative apocalypse feared by traditionalists. For her, it’s a personal journey that requires artistic vision, technical patience, and thoughtful consideration of how to share authentic expression in a complicated landscape.

Her months-long song-making sessions, careful attention to emotional authenticity, and wrestling with ethical implications aren’t the actions of someone lazily automating creativity. They’re the practices of an artist using new tools for the oldest human reasons: to process experience, find joy, and connect with others.

Back in Canada now, Ola continues creating music that emerges from genuine personal need and curiosity. She’s got a backlog of inspiring songs to release but she’s only willing to put them out when she’s ready, and with intention: this music serves her first, and if it helps others, that’s a beautiful bonus.

“We need to be gentle and make it fun,” she explains of her approach to creating positive content. “Finding really wonderful ways to help people feel better, I think is the way to go, and music is the way to do it.”

You can find Ola on all the major streaming platforms including Spotify, Apple Music, and YouTube, as well as on Bluesky.