Technically this is already possible, but we force the user to select a language to reduce the error rate and increase the quality of the speech recognition.

There could be a case though where a user really needs to switch languages frequently during a meeting, so the upsides of having that option could outweigh the downsides.