When David Betts was diagnosed with ALS, he became fixated on finding a tool that could help him keep his voice – but none of them sounded human.
“I didn’t know how long I would have my voice,” Betts told Technical.ly. “I faced this uncertain deadline.”
Betts, a native Pittsburgher, has always processed the world out loud, he said, solving problems by speaking about them, but his ALS diagnosis threatened that. The degenerative disease breaks down muscle control, which can affect a person’s ability to speak. When that happens, patients often turn to assistive tools to help them communicate, but Betts couldn’t find one that wasn’t robotic.
“The feedback that has stuck with me? ‘I want to hear daddy read that story a hundred, hundred times again, forever.’”
David Betts, Talk To Me, Goose!
“A pivotal moment for me was hearing another ALS patient talk about the ‘awkward pause’ in conversation, that long silence while you type, while people wait, while the conversation slowly moves on without you,” Betts said. “I realized that with all our advancements in artificial intelligence, we should be able to eliminate that gap.”
Betts made it his mission to get that tech into the hands of more people. He created Talk To Me, Goose!, an AI-powered text-to-speech app that helps people who are losing their ability to speak to communicate more naturally. It uses voice cloning, or a digital replica of how someone sounds, to recreate users’ voices, breaking the mold of the tech’s reputation as a tool primarily used for predatory scams.
Launched last year, the app already has hundreds of users and is gaining global recognition for its human-centered approach to accessible technology.
“Communication should be a right not a privilege,” Betts said. “I don’t see a need to make any more barriers for people living with this disease. They already have too many,” he added on why he offers the app for free to people with ALS.
A former senior partner at Deloitte, Betts had no technical coding skills before he was diagnosed with ALS. After a few weeks of taking online classes and spending many late nights studying tutorials, Betts launched Talk To Me, Goose! (affectionately named after the character Goose from the film Top Gun) in March 2025.
Users can create a voice clone with as little as 45 seconds of audio, according to Betts, and the app also works with eye-gaze devices, making it accessible to users with limited mobility.
The app’s AI assistant, Merlin, turns short text inputs into full sentences, helping reduce the long, awkward pauses that are common with traditional text-to-speech tools. For example, a short prompt of “I’m cold” can quickly be transformed into “Hey, when you have a moment, I’m starting to feel a little bit chilly, could you grab me a blanket, please?” to better match the tone the user would like to use.
“I didn’t want to communicate as a generic machine when technology was already capable of better,” Betts said, “and I didn’t want anyone else to have to settle for that either.”
Challenging AI’s contentious reputation
Artificial intelligence is increasingly being integrated into assistive technology — tools designed to help people with disabilities communicate or navigate the world with more ease — even as concerns over bias, data privacy or environmental impact persist.
For Betts, criticism around his use of AI has come as a surprise.
“I have at times been taken aback by how negative some of the feedback has been from maybe outside the disability community in particular,” he said.
While critics may say AI-powered tools risk replacing human work, mishandling sensitive data or causing a negative environmental impact, Betts instead argues Talk To Me, Goose! is a worthwhile use case for the tech.
It’s an approach that Big Tech has also embraced. Companies like Apple and Google have released AI-powered accessibility tools aimed at making daily life easier for people who are visually or hearing-impaired.
In recent years, text-to-speech advancements, especially with AI becoming more mainstream, have focused on “style preservation,” according to Carnegie Mellon University (CMU) PhD student Karen Rosero. That means focusing on cloning voices and maintaining an individual’s speaking style, which can be especially helpful beyond Betts’ use case in fields like speech-language pathology.
Researchers at CMU have recently developed a new AI tool that helps kids with speech disorders. Using machine learning, it analyzes audio clips of mispronounced words and transforms them into corrected speech, letting children hear the proper pronunciation in their own voice.
Rosero and CMU professor Carlos Busso are now working to integrate a visual element into the tool, all while keeping accessibility and ethics as a focal point, according to Rosero.
“We are not training our AI systems with patient data,” Rosero said, in order to protect children’s privacy. The goal of the project is not to replace speech language pathologists but to support them, since there’s a national shortage, she said.
For Rosero, the project also hits close to home. About a decade ago, her grandmother died from ALS, and she witnessed firsthand the gradual loss of her grandmother’s ability to move and speak.
“I didn’t just want to build AI models that can be used for big industries,” Rosero said, “but instead something with an accessibility component to it, bringing these technologies closer such that everyone can use them.”
Building a voice for those who’ve lost theirs
Since launching his app, Betts has formed a partnership with the Live Like Lou Foundation, making the tool free in the US and Canada for people living with ALS.
Next month, Betts will travel to the United Nations in Vienna to present Talk To Me, Goose! at the Zero Project Conference as a 2026 awardee. The global initiative works to advance a UN treaty to improve the quality of life for people with disabilities, and its annual awards recognize tech that removes barriers to communication and accessibility.
“To be awarded, I mean, it’s really a little overwhelming, to be honest,” Betts said. “This is a solution that started from nothing on [Jan. 5, 2025] to 650-some odd users in less than a year.”
At the Vienna conference, Betts said he hopes to form more partnerships that can make the app free for the 97 million people around the world who live with speech disabilities.
“I don’t have time to waste,” he said. “I take inventory as to what works and doesn’t every day and there’s a little less than there was the day before. That’s why I’m working as fast as I can right now, so in the worst case, this stimulates the existing providers to innovate.”
Betts recently released a spinout app called Fable’s Adventures, which allows users to quickly generate custom stories and turn them into audio using natural‑sounding AI voices, helping families who use Talk To Me, Goose! tell bedtime stories.
For Betts, those small moments are the ultimate measure of success.
“The feedback that has stuck with me?” Betts said. “‘I want to hear daddy read that story a hundred, hundred times again, forever.’”
