UG, Google, and UCL develop speech recognition technology for Ghanaian languages 

Nii Martey M. Botchway  

Accra, Nov 8, GNA – The University of Ghana (UG), in collaboration with Google Research Africa and the University College of London’s (UCL) Global Disability Innovation (GDI), has launched an advanced speech recognition project tailored to Ghanaian languages.  

The initiative, named Tɛkyerɛma Pa, aims to enhance communication for individuals with non-standard speech patterns due to conditions such as Cerebral Palsy, Stroke, Down Syndrome, Amyotrophic Lateral Sclerosis (ALS), and Parkinson’s disease. 

The project focuses on improving AI-powered speech recognition for five key Ghanaian languages—Akan, Ewe, Dagaare, Dagbani, and Ikposo, to provide better accessibility for people with speech impairments, making it easier for them to communicate and access technology. 

Speaking at a media briefing in Accra, Professor Isaac Wiafe, Project Lead and Associate Professor at the University of Ghana, shared updates on the project’s progress. 

He highlighted the importance of the initiative in advancing inclusive communication for all individuals, regardless of speech challenges.  

According to Prof Wiafe, existing speech recognition technologies have limitations when dealing with non-standard speech patterns, which often resulted in inaccurate interpretations. 

To address this, Tɛkyerɛma Pa is training AI models to better understand these unique speech patterns in the local languages.  

The project’s long-term goal is to improve the accessibility and effectiveness of speech recognition systems, thereby empowering Ghanaians with speech disabilities and helping them communicate more effectively. 

“By developing localised AI solutions, we can ensure that more Ghanaians have the opportunity to communicate effectively, regardless of their speech patterns. 

“We are thrilled to lead this important initiative, which reflects our commitment to using technology to solve real-world challenges,” Prof. Wiafe said.  

Ms. Katrin Tomanek, a Senior Research Scientist at Google Research, said that the Tɛkyerɛma Pa project builds on continuous research aimed at making Automatic Speech Recognition (ASR) more accessible.  

She explained that the initiative was designed to provide valuable insights that would help local researchers create ASR models tailored for Ghanaian and other African languages.  

This, she said, aligned with Google’s ongoing commitment to advancing AI research in Africa and ensuring that technological progress benefits all individuals, regardless of their speech patterns. 

Prof. Cathy Holloway, the Academic Director at the GDI Hub, emphasized the importance of extending AI-powered recognition to more languages to help people with non-standard speech communicate more effectively.  

She noted that the project would serve as a foundational step for future language models, with the potential to benefit individuals not just in Ghana, but also contribute to global AI development. 

Prof. Holloway further explained that Google’s vision is to support local researchers in leading innovations that address both local and global challenges by providing funding, technical expertise, and cloud computing resources. 

“The initiative forms part of Google’s broader commitment to advancing AI development in Africa,” she said. 

The Tɛkyerɛma Pa project is supported by a $40,000 grant from Google, with research expertise provided through the AT2030 program, an assistive technology initiative from the UCL’s GDI Hub. 

GNA