A Singer Lost His Guitar to Parkinson’s. He Hummed His Way Back to Music

Samuel Smith used to shape songs with a guitar in his hands. Then the tremors started. Diagnosed with Parkinson’s disease in 2020, the London singer-songwriter watched his ability to play slip away. Stiffness. Fatigue. The precise finger work that defined his Americana sound faded.

But he finished a new album anyway. “The Art of Letting Go,” released this year, includes eight tracks built with help from artificial intelligence. For the instrumental “Horizon,” Smith hummed rough melodies into his phone. He fed those recordings into AI tools like Suno and Udio. The systems generated synthetic demos. Session musicians in Nashville used those demos to understand what he wanted. The AI tracks never made the final cut. They just translated his vision when his hands couldn’t.

“Don’t play, don’t be creative, or find a way out, find a route,” Smith, 49, told Fortune. “And for me, this was the route.”

The process took persistence. Smith often needed 50, 100 or even 150 attempts to nail down a single demo. He spent hours editing. He fed the systems his lyrics and detailed notes on instrumentation, mood and style. The output acted as a bridge. Real musicians—including Grammy winners Jerry Douglas on dobro and Alison Brown on banjo—brought the tracks to life under producer Matt Rollings.

Smith released his debut, “In the Springtime,” in 2023. He wanted his young sons to hear him perform while he still could. The second record carries heavier stakes. His four-year-old may never remember him playing guitar. The disease, progressive and unforgiving, had already shaken his core identity as a musician. “I’d always written, I’d also played, I always sung,” he said. “And immediately it became clear to me that I was in trouble, that my music was going to be seriously compromised.”

Yet one moment on the new album captured something irreplaceable. Grammy-nominated guitarist Julian Lage joined for the title track and “Horizon.” Smith managed a brief guitar duet. “I hadn’t been able to play for months, but I kept telling myself that if I wrote something to take to the studio, perhaps the clouds would part for a few minutes,” he recalled. “That’s what happened. I had a window of about 10 minutes in the studio when my arm freed up. So in the end, I was able to capture the last breath of my guitar playing.”

His story lands amid fierce debate over generative AI in music. Sony Music Entertainment, Universal Music Group and Warner Records sued Suno and Udio in 2024, accusing them of training on copyrighted material. Settlements followed with some labels. Critics, including artists like Tift Merritt and David Lowery, warned in an open letter that AI-generated tracks could dilute royalties and enable fraud. Smith sees a different side. “AI is not replacing anything for me,” he said. “It’s unlocking, it’s enabling. It’s allowing me to keep writing. I upload my lyrics; AI doesn’t create my lyrics. I upload my music; AI does not create my music. It then brings it to life in a way that I can play to session players and say, ‘Here, that’s what I’m thinking, that is what I’m hearing.’”

Experts point to broader potential. Ruaidhri Mannion, a composer and professor at Brunel University of London, notes that earlier technologies like affordable digital recording already widened access. AI prompt-based tools could extend that further for people facing physical barriers. “If these tools are able to enable people to be able to participate with other creative groups and encourage more people to feel confident to be able to reach out to an ensemble or an orchestra or something, then I think that is all for the better,” Mannion told Fortune. He cautions, however, that overreliance risks bypassing the friction essential to artistic growth. Collaboration, failure and iteration matter.

Smith’s experience adds to a surge of AI applications in Parkinson’s care. Voice analysis models now detect the disease with striking accuracy by examining pitch, rhythm and articulation. One study achieved 99% accuracy in early identification. Another, published in Scientific Reports, reached 91% accuracy with explainable AI techniques that highlight which vocal features drive predictions. A Toronto teenager built a tool that analyzes just three seconds of speech. These methods promise faster, cheaper screening than traditional clinical exams.

Treatment has advanced too. In February 2025 the FDA approved adaptive deep brain stimulation. The system uses AI to adjust electrical pulses in real time based on a patient’s brain signals. Conventional DBS delivers constant stimulation. The adaptive version responds to daily fluctuations. Neurologist Fernando Pagan at MedStar Georgetown University Hospital called it a long-awaited step. “We’ve been waiting for this, to be honest,” he said in a May report from InvestigateTV. Patient Rick Schena, who received the upgrade in 2025, reduced his daily medication by half. He regained the ability to fly after 18 years. “Anything that’s going to make my life better, I’m all for,” Schena said.

Researchers at the Michael J. Fox Foundation have applied AI to massive datasets from the Parkinson’s Progression Markers Initiative. The models spot subtle patterns in imaging, genetics and sensor data that humans miss. They identify biological subtypes, forecast symptom progression and accelerate drug discovery. Early detection remains the priority. Movement analysis from simple video of finger-tapping tests already reveals changes invisible to the naked eye, according to work at the University of Florida.

Smith took his insights public. On May 21 he joined the Berklee Music and Health Institute in New York. The event gathered industry leaders, researchers and clinicians to explore music’s role in neurological conditions. He performed with musicians from his album. The message he delivers to AI developers is direct. “My message would be that if these companies want to show they’ve got a place, a role in society, then step up. Engage with health professionals, engage with music therapists, engage with society and show us what you can do.”

Creating music now looks different for him. No more solitary guitar sessions in the same way. Instead he works from a home studio, humming ideas, refining prompts, iterating with software. The frustration has not vanished. Parkinson’s continues its advance. But the album stands as proof of adaptation. “It is no exaggeration to say that AI saved my songwriting,” Smith said in a release tied to the record. “I wouldn’t have completed this record without it.”

His two sons may hear that legacy differently. The older one remembers the guitarist. The younger may know only the songwriter who refused to let the disease define the final chapter. In that refusal lies a larger point. Tools once viewed solely through the lens of industry disruption can restore agency to those losing physical capacity. The debate over copyright and royalties will persist. So will questions about artistic authenticity. Smith’s case suggests another dimension. When a hand can no longer fret a chord, a hummed melody processed through algorithms can still reach session players across an ocean. The music arrives anyway.

Recent coverage reinforces the momentum. A May 6 InvestigateTV report detailed how adaptive stimulation reduces medication needs and side effects while improving motor control. Studies on voice biomarkers continue to refine accuracy and add explainability, making clinicians more likely to trust the outputs. Foundations and universities pour resources into multimodal AI that fuses speech, gait and handwriting data. The goal is not replacement of human judgment but augmentation of it. Early intervention, personalized therapy, better quality of life. These aims align with what Smith experienced in the studio. A brief window opened. He played. The recording preserved it.

Parkinson’s affects more than 8.5 million people worldwide. Numbers are rising. Any method that eases the burden, whether through earlier diagnosis via a smartphone recording or creative continuity through generative audio, carries weight. Smith understands the perils of the technology. He also lives its possibilities. His album is out. The guitars on it belong to others. One short duet remains his own. The rest came from humming into a phone, feeding prompts, iterating until the machines spoke his language well enough for humans to follow. The route worked.