London Singer Overcomes Parkinson's to Complete New Album with AI

London-based singer-songwriter Samuel Smith, who was diagnosed with Parkinson's disease in 2020, has seen his guitar-playing ability gradually stripped away by the illness. With the help of AI music generation tools, he has finally completed his second album by humming melodies to convey his musical vision to other musicians.

According to the Associated Press, Smith had spent years writing songs with a guitar in hand. Now, with Parkinson's significantly limiting his playing ability, he has turned to artificial intelligence (AI) tools to maintain his creative output. His story has drawn attention to the potential of AI in assisting disabled musicians and the controversies surrounding it.

Smith recently released his second album, "The Art of Letting Go." Among the eight tracks is an instrumental piece titled "Horizon." He used an AI music generation platform to create demo arrangements, which allowed him to communicate his creative vision and concepts to the musicians responsible for the final recording.

Smith emphasized that the demos he created by humming rough melodies into his phone and uploading them to song generators like Suno and Udio were absolutely not intended to be mixed into the final studio version of "Horizon."

49-year-old Smith says that during the year-long album production process, common symptoms of Parkinson's—tremors, stiffness, and fatigue—caused his guitar skills to continue to decline. For him, this was the way out.

Generative AI has sparked controversy in the music industry, with artists and record labels complaining that their copyrighted works are being used to train the models behind AI music tools.

Sony Music Entertainment, Universal Music Group, and Warner Records filed lawsuits against AI music generation tools Suno and Udio in June 2024, although some companies have since reached settlement and partnership agreements.

Less discussed is what these platforms can do for professional musicians like Smith, whose illness has affected his hands. "The condition made me realize immediately that I was in trouble, and that my music would be far from what I expected."

AI music generators use massive datasets of recordings and audio to train their systems. These platforms analyze melody, harmony, and rhythm patterns, then generate new audio based on prompts or uploaded recordings. Users don't need musical talent to get a decent, or even popular, song.

After Smith hums a song into his phone and uploads it, he provides prompts describing instruments, mood, and style. He says that creating a satisfactory demo from the tracks generated by the app often requires "50, 100, 150 attempts" and a lot of editing "to get something that sounds close to my music."

"AI isn't replacing anything for me," he says. "It's unlocking, it's empowering, it's allowing me to continue writing songs. I upload my lyrics; the AI doesn't write my lyrics. I upload my music; the AI doesn't write my music."

Smith adds: "Then, it presents this music in a living way, allowing me to play it for the collaborating musicians and say, 'Look, this is the concept in my head, this is the sound I hear in my ears.'"

Ruaidhri Mannion, a composer who teaches at Brunel University of London, says that technologies like affordable digital recording software have "effectively democratized music creation" in recent decades.

Mannion points out that these AI tools, which can turn "voice or text prompts" into "sophisticated, mature-sounding material," make it easier for songwriters and musicians to communicate inspiration and collaborate.

However, he also notes that trial and error, frustration, and refinement are indispensable parts of achieving art. Over-reliance on technology could interfere with these processes.

Udio and Suno deny copyright infringement and say they hope to work with the music industry rather than against it. However, some musicians are not convinced. A group of musicians and activists published an open letter in February titled "No to Suno."

The open letter states, "Many in our community are embracing responsible AI as a creative tool, allowing fans to explore and interact with the art we create, which is great."

"But this is very different from creating an environment where AI-generated works derived from our music are circulated in bulk to dilute our royalties. Even worse, it rewards those who actively seek to defraud. Artists need to understand the difference between the two."

Smith says he believes his personal experience demonstrates how AI, when developed responsibly, can benefit society and expand creative channels.

Smith stated that his message is that if these companies want to show they can contribute to society, "then step up and work with medical professionals, work with music therapists, work with all sectors of society, and show us what you can do."