Key takeaways:
Voice banking uses artificial intelligence (AI) to record a computerized version of someone’s voice. This helps them communicate if they can’t speak naturally.
The process involves recording the person’s speech and creating a unique synthetic voice.
Voice banking is particularly beneficial for people with conditions like amyotrophic lateral sclerosis (ALS) or other speech or language disorders.
Jim O’Brien learned about voice banking soon after he was diagnosed with amyotrophic lateral sclerosis (ALS) in April 2022. But he didn’t rush to preserve his voice using artificial intelligence (AI).
“When I first heard, I didn’t think much of it. I understand being proactive and staying one step ahead. But I had all this other stuff going on,” says Jim, a 63-year-old father of four, grandfather of six, and great-grandfather of one. “But as I continued on my journey, it became apparent very quickly that voice banking was something to be proactive about.”
ALS is a progressive neurodegenerative disease that affects nerve cells in the brain and spinal cord. As ALS progresses, it can lead to the gradual loss of muscle control and function throughout the body. One of the common symptoms of ALS is the loss of speech or difficulty speaking, also known as dysarthria.
Someone in one of Jim's online ALS support groups made a comment that hit home. The comment was from someone who worried about eventually being unable to speak normally to their grandchildren. They would need to rely on synthetic voice for communication.
“That really struck a chord with me,” says Jim, who served in the U.S. Air Force for 6 years, then worked in local law enforcement near his home in Des Moines, Iowa. “There’s no harm in doing it because you don’t have to use it. But if you choose not to bank it and you do lose your voice, you can’t go back and do it later.”
Voice banking uses AI to “bank” someone’s voice. It has come a long way in the last decade. Five years ago, those seeking to preserve the sound of their natural voice often needed to record themselves talking for hours. The cost could be prohibitive.
But now that more companies have gotten involved with voice banking, recording your voice can be done relatively quickly. There are also organizations, like the ALS-advocacy group Team Gleason, which can provide financial help.
Another sign of how far voice banking technology has come: Apple will soon be releasing its own voice-replicating software. Personal Voice, available in late 2023, is software that allows users to create digitized versions of their voices in as few as 15 minutes using an iPhone, iPad, or MacBook.
The ALS Association encourages people diagnosed with ALS to bank their voices as soon as they can. As its website notes: “People with ALS often lose their ability to speak. For some, this change happens quickly; for others, this change may be more gradual. In any case, it’s important for people with ALS to consider voice banking and/or message banking as soon as possible after diagnosis, while they still have their strongest voice possible.”
Jim used Acapela Group software to digitize his voice — which sounds like this. When the company began offering the service in 2015, users needed to say 350 sentences aloud for their voice to be preserved. Since 2020, the company has only needed 50 sentences.
Jim's voice becomes scratchier and weaker as the day progresses.
“It’s like you speak, and not as much comes out as you’d like. There are times when it will cut out a little bit,” he says. He needed multiple days to record his 50 sentences.
The software rejects sentences that aren’t spoken clearly. Jim says it asks users to repeat random sentences, such as: “I took the glass and set it on the counter.”. It collects a variety of sounds the AI program needs to replicate his voice.
Jim says he was shocked by how closely it matched his natural voice.
“It was very much like my voice minus some inflection and things like that,” he says. “But you wouldn’t know it’s a synthetic voice.”
Jim also recorded himself saying more personal words and sentences using “message banking.” Message banking is a process where users record themselves saying words or phrases that are unique or special to them and then have the option of playing the recording when desired.
A roundup of oft-recorded message banking phrases compiled by Boston’s Children’s Hospital includes personal requests such as, “I want a kiss,” and “Leave me alone,” as well as humorous quips such as, “That’s easy for you to say,” and “That was the voice machine talking, not me!”
For Jim, messaging banking included “things I might say to my dog, my nicknames for [my] grandchildren, and the real personal type things. There were a few curse words as well,” he says. He also recorded a personal catchphrase, “I call bull— on that.”
Jim's keyboard has shortcuts that help him quickly access these recordings.
“I have a button that says ‘pets.’ And when I hit the pet button, it gives me a choice of being all goo goo ga ga for my dog or my sister’s dog,” he says. “It really adds a personal touch.”
Because Jim expects to lose the use of his hands, he’s begun practicing with eye gaze technology to type messages and push buttons. It allows users to use a computer by looking at words or commands on a screen.
“As my journey with ALS goes on, that could be my only means of communication with the world around me,” he says. “When you boil it down, that’s priceless.”