Google Research Africa, in collaboration with the University of Ghana and UCL’s Global Disability Innovation Hub (GDI Hub), has unveiled Project Euphonia, a transformative initiative aimed at enhancing speech recognition technology for individuals with non-standard speech. The project, which focuses on Ghanaian languages, is designed to make artificial intelligence (AI) more inclusive and accessible.
Backed by a $40,000 grant from Google, the initiative seeks to address the shortcomings of conventional automatic speech recognition (ASR) systems, which often fail to accurately interpret diverse speech patterns. By prioritizing local languages and linguistic nuances in Ghana, the collaboration aims to create ASR models that empower individuals with speech impairments to communicate more effectively.
The project combines Google’s advanced technical expertise, the University of Ghana’s research capacity, and the GDI Hub’s focus on inclusive design under its AT2030 program. Together, they are working to develop the first-ever open-source dataset of non-standard Ghanaian speech, an essential resource for fostering AI inclusivity in Africa.
“This partnership leverages local insights to build AI models that better understand diverse speech patterns,” said Katrin Tomanek, Senior Research Scientist at Google Research. “The open-source dataset will be a milestone for African AI development and a step toward making technology accessible for all.”
Professor Isaac Wiafe, Associate Professor at the University of Ghana and Project Lead emphasized the initiative’s significance for local communities. “This project demonstrates our commitment to solving societal challenges with technology. By focusing on Ghanaian languages, we’re ensuring inclusivity in the digital era,” he remarked.
The initiative builds on Google’s broader commitment to fostering AI innovation across Africa. It complements tools like the Project Relate app, which assists individuals with speech impairments worldwide by improving communication and accessibility.
Professor Cathy Holloway, Academic Director at GDI Hub, highlighted the global potential of the project: “This collaboration provides a blueprint for expanding AI-driven speech recognition to underserved languages and regions. It sets the stage for similar efforts worldwide, ensuring accessibility for all.”
As Project Euphonia progresses, its leaders hope to expand the initiative to include more languages and regions, ensuring the transformative power of inclusive AI reaches communities across the globe.
Article by Nyokabi Wanjiku
Photo/Google
Comment