Audio interviews are currently broken. People can use AI and many will do. Not necessarily for speech generation but to know what to say.

For research studies, we slowly revert to on premise physical interviews at work. If we want the ChatGPT answers, we don’t need another human in the loop.