How this startup uses AI to give speech-impaired people their voice back

There is big tech — and then there is smaller tech with big impact. Dutch AI startup Whispp is using assistive voice technology to convert speech diminished by ailments such as stuttering, throat cancer, vocal cord paralysis, and ALS, into the person’s original, natural voice.

The startup, based in Eindhoven, just secured a €750,000 seed funding round led by LUMO Labs to launch its assistive voice technology and its first paid and subscription-based calling app. Whispp will also use the funds raised to expand into European and US markets. Furthermore, the company is one of this year’s CES 2024 Innovation Award honourees. 

Voice distortions from stuttering or following afflictions such as throat cancer or ALS affect over 300 million people worldwide. Whispp’s proprietary technology uses recordings of a person’s “old” healthy voice to turn their real-time, affected speech into a clear voice. This is then modified with the intended intonation, modulation, and emotion.

Those who still have a natural voice but for instance stutter, can simply record their current speech with the app. 

This audio-to-audio-based approach eliminates the typical latency of 2-3 seconds for text-to-speech technologies, which can create barriers and disrupt communication. The technology also builds on the fact that for several voice and speech disorders it is helpful to deliberately steer the voice towards whispering, due to neurological changes in the speech system.

“Our big bold dream is to have Whispp’s assistive voice technology available on every smartphone and laptop worldwide to create a more inclusive world,” said Whispp founder and CEO Joris Castermans, adding that LUMO Labs’ experience would help the company partner with network operators and tech companies to help achieve its goal. 

AI voice tech for good

Castermans, who himself stuttered as a child, founded Whispp together with Akash Raj, who is the company’s CTO, in 2019. The latest funding follows a 2022 pre-seed round led by Libertatis Ergo Holding B.V., an independent, wholly-owned subsidiary of Leiden University, which also participated in the seed round. 

Essentially, this is a perfect example of how tech isn’t good or bad in and of itself — its impact will depend on how we use it. As the world frets over deepfakes (and with good reason), we can also use AI voice technology to enhance human communication and connection. 

It could also significantly contribute to diversity, equity, and inclusion. 

“The ability to speak heavily affects an individual’s wellbeing and sense of belonging. More than 1% of the global population however suffer some kind of progressive loss of speech and/or speech impediments,” LUMO Labs founding partner Andy Lürling stated. “Whispp allows these people to literally be heard and participate and contribute to society like anyone else.”