Casey gets his voice back

March 25, 2026·37 min·3 clips

Casey Harrell, paralyzed by ALS, now talks through 256 electrodes in his brain that decode his thoughts into speech.

Host Julia Longoria opens on a video call from Miami, struggling to start an interview with climate activist Casey Harrell in Oakland. Harrell was diagnosed with ALS in 2020 and lost the ability to speak within three years. In 2023, UC Davis researchers implanted 256 electrodes in his motor cortex as part of a brain-to-speech interface that trains an AI on his speech-related brain activity. The system is reported as 97 percent accurate and is good enough for Harrell to use on his own, a notable contrast with older eye-tracking systems used by people like Stephen Hawking. Longoria describes the rhythm of the interview, with long silences while Harrell corrects misrecognised words using eye gaze on a second screen before the AI voice speaks his sentences aloud. Harrell uses a voice synthesised from his old public speeches. He explains his disease started in his legs and arms, then hit his voice about two months after diagnosis, fading gradually over a year. He calls losing his voice the worst part of ALS and names Raffi's Baby Beluga as the song he most wants to sing again. He tells a story of his young daughter standing up in a speech therapist's office to translate for him when the specialist could not understand him. He also describes coughing attacks, a hallmark of ALS respiratory decline. In the second half, his wife LaVonna Saxon recounts meeting Harrell through Rainforest Action Network and the slow, brutal loss of his speech, including years when basic caregiving was dangerous because she could not understand what he needed. She describes learning through a cousin about UC Davis's trial, the implant surgery, and the research team's low expectations. At the first session, the system worked immediately, and Harrell used his first synthesized speech to read his wedding vows to her in front of a film crew. The family now has a privacy button that prevents research capture. Saxon notes the interface also works while Harrell is in deep sleep, producing words that match his dreams, and calls for dream researchers to study it with the BrainGate team. She ends by reflecting that she no longer dreams because she is woken so often as his caregiver.

As heard by us

A patient, clear-eyed look at brain-to-text AI and the effort behind Casey Harrell being heard.

Casey Harrell's halting conversation with Julia Longoria turns a brain-to-text AI interface into something more complicated than a breakthrough demo.

Read the full review in PlayNext →

Why you'd press play

You want to hear a voice fought for, not just found.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Unexplainable