Advancing Speech Technology for All

The Centre for Digital Language Inclusion (CDLI) is leading the way in making automatic speech recognition (ASR) work for people with non-standard speech, especially in underrepresented African languages.

This new report highlights:

• Pioneering AI for Inclusion: CDLI is developing ASR models for 10 African languages, collecting speech data from individuals with cerebral palsy, stammering, and other speech differences.

• Community-Led Innovation: A practical “cookbook” approach empowers local developers to build speech tools for their own communities using open-source methods.

• Key Research Findings: Includes evaluations of Google’s Project Relate in Ghana and studies on ASR use by people with ALS, showing both promise and the need for localization.

At its core, this work is about communication equity, ensuring every voice can be heard, no matter the language or ability.