Jules Rodriguez, who lost his speech to ALS, built an AI voice clone from old recordings in June 2024, giving him a realistic version of his pre-illness voice for daily communication.
The clone captured breaths, cadence and stutters that older voice-banking systems often missed, and Rodriguez said hearing it made him feel he had "reappeared"; his wife cried when she first heard it.
That restored voice quickly became practical as well as emotional: after a tracheostomy on Oct. 1 permanently ended his natural speech, Rodriguez used eye-tracking text input and the clone to keep talking.
By mid-November, the same technology had helped him resume a long-abandoned goal of performing stand-up comedy, while other ALS users cited similar gains in singing, reading to family and social connection.
Clinicians and nonprofit advocates said the newer AI voices can be made from archived audio rather than 350 to 1,600 scripted phrases, and in some cases have even influenced patients' end-of-life decisions.