UMC Utrecht develops AI voice for people without vocal cords
A patient and healthcare professionals are working on a prototype designed to make a voice prosthesis sound more personal.
UMC Utrecht is working on a prototype that uses artificial intelligence to turn the sound of a voice prosthesis into a recognisable personal voice. The project is still in development and is not a proven or generally available treatment.
The idea came from Boudewijn, a patient who lost his vocal cords and larynx after laryngeal cancer. A voice prosthesis allows him to produce sound again, but he says it sounds different from his original voice. He therefore suggested using old voice recordings for a digital reconstruction.
Together with head and neck surgeon Weibel Braunius and innovation manager Mart Schoondermark, Boudewijn is working on the ReVoice project at UMC Utrecht. The aim is for the technology to convert the sound of the voice prosthesis, while someone is speaking, into a voice that more closely resembles the user's original voice.
The technology is still a prototype. There is no evidence yet that the system is reliable, safe and suitable for large groups in everyday healthcare. Nor is it clear how well a voice can be reconstructed when someone has few usable old recordings or when their way of speaking changes significantly.
A total laryngectomy has major consequences for communication and daily life. People have to learn to speak with a prosthesis and receive support from healthcare professionals, including in speech therapy and rehabilitation. The precise options vary from patient to patient and depending on the treatment.
The project will be in the final of UMC Utrecht's Ureka Mega Challenge in October. The winner can receive 50,000 euros to develop the idea further. That money could be used for technical development, research into usability and involving more patients.
The potential benefit of a personal AI voice is not limited to intelligibility. A voice is also connected to identity, social contacts and the way someone experiences themselves. At the same time, future applications will need to handle voice recordings and consent carefully, as well as the risk that a system imitates someone's voice incorrectly or without permission.
Fact-check Approved · Nour Haddad — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
The development, UMC Utrecht's involvement and the prototype status were checked directly with the hospital. Medical context has been kept limited and no individual treatment advice has been included.
- confirmed UMC Utrecht is developing, through ReVoice, an AI prototype for a more personal voice after laryngectomy. — Described in UMC Utrecht's patient and innovation story. source
- confirmed A voice prosthesis makes speaking possible after removal of the larynx. — UMC Utrecht's patient information describes how a voice prosthesis works. source
- confirmed The ReVoice project is not yet a generally available treatment. — UMC Utrecht describes the system as a prototype that requires further development. source
Editor's note
The ReVoice project and its status as a prototype have been confirmed by UMC Utrecht. The technology has not yet been clinically validated; the article makes no statement about individual treatment choices.Sources
More on this in Dutch media
- NU.nl — „stemprothese hoofd-halskanker”
- De Telegraaf — „stemprothese hoofd-halskanker”
- AD — „stemprothese hoofd-halskanker”