A Manx PhD student has created the first artificial intelligence system capable of speaking and translating the Manx language.

Chris Bartley, a researcher in computer science specialising in language and speech technologies, has developed a text-to-speech programme as part his PhD research that converts written Manx into spoken audio using a synthetic voice.

Text-to-speech technology is widely used in major languages and underpins tools such as virtual assistants, navigation systems and screen readers - but these systems rely on large amounts of training data to learn how a language sounds.

Talking about his project, Chris said: ‘The reason why Manx has never had this previously is because the systems require training data to be able to work.

‘For languages like English, German and all the well known European languages, it's quite easy to find that data. It's very much not the case for Manx.’

Despite these challenges, the system has already reached a stage where it can be used for practical applications. One of the most immediate uses is screen reading, which allows written text to be read aloud and can significantly improve accessibility for visually impaired users.

‘As it is, I think it's definitely good enough for certain applications. It all depends on what we want to use it for,’ Chris added.

‘For something like screen reading, I think the system that we have is currently good enough. Once I improve the systems, my long term goal is to have it in an educational setting of some sort.’

Chris has been in contact with Manx language organisations, including Culture Vannin, and with those involved in the technical development of the Manx corpus (https://corpus.gaelg.im/), to explore potential collaborations and applications.

A demonstration of the AI system speaking Manx can be found by visiting https://www.youtube.com/shorts/3GyQQlM79bc