I have an Alexa skill that was originally written back when US English was the only dialect of English supported.
Since then, the skill has been automatically migrated to provide support for Canadian, UK, Indian, and Australian English. Now, we're going in to update the skill and we've noticed that each of these localisations has its own independent language model.
Since our skill is for a local institution within the US, it seems like overkill to maintain five versions of the language model at our scale. Is it possible to remove localisations from a previously deployed Alexa skill, so we can maintain a single copy and not need to worry about the other four localisations getting out of date? We have no intention of maintaining these auto-generated localisations, they'll just turn into technical debt as time goes on.