Hi-Tech Tools for Lo-Literacy Learners: Using AI accent recognition to enable adults to improve their literacy
Virtual assistant apps are a good supportive method to help adult learners improve their literacy outside of a classroom environment. Users can read out written words as well as spell out words that are played to them in audio snippets and be assessed by the app. However, this interaction relies on the app being able to recognise the words the learner is saying regardless of their accent or pronunciation. Something which is even harder with short words or phrases. In the UK, there is a wide range of different languages, accents and dialects spoken across the country. And in Scotland, where the project is based, the concentration of regional accents is greater still, making speech recognition more difficult. This can have a negative effect on people’s ability to learn by reducing their confidence and potentially adding to a sense of discrimation or underrepresentation and compounding inequalities.
City of Glasgow College’s idea is for each user to help to train the App on their specific voice during the setup process to create a personalised accent profile from which the App can then recognise the rest of the word library. Using Google’s Teachable Machine platform, which helps users design machine learning programmes, the App’s voice analysis will learn the users’ voices. This means they can use the literacy app without their natural accent being an issue, helping them to learn more effectively.
City of Glasgow College and its partners have already created the basic app, “Citizen Literacy”, with its supportive features and have conducted testing of the accent recognition AI interface which has shown early promise.
John Casey, Project Manager and Learning Designer at City of Glasgow College, says: “The AI for the Citizen Literacy partnership has the potential to transform our voice recognition service in our literacy programme. Being able to include regional accents that currently are not well served by standard voice recognition systems means none of our target learners will be excluded from using the cutting edge learning systems we are currently designing to improve adult literacy in Scotland.”
This project is led by John Casey, Learning Designer at City of Glasgow College in partnership with Mirco-phonics Software, ReachWill and Bolton College. The project has received funding from Nesta in Scotland’s AI for Good programme and Nesta’s CareerTech Challenge Prize in partnership with the UK Department of Education.