I think a great use case for this technology could be to preserve dying languages. I'm sure a lot of work has already gone into preserving the written form of t

I think a great use case for this technology could be to preserve dying languages. I'm sure a lot of work has already gone into preserving the written form of these languages, but training models on data sets of native speakers could be a way to preserve pronunciation.