Many of these tools go between written and spoken language media types. Sometimes, this means generating spoken language from a text-based file, and sometimes, this means going in the other direction.
Other kinds of audio-focused generative AI focus less on language sounds, and more on a category of audio information (music, for example). As with generative AI tools developed around visual media, there is great potential within this family of tools for improving the accessibility of information contained in digital objects.