How do you use ai to record audio? I mean, recording audio is just procedural code. You can use ai to process the recording, of course.