How were you communicating with the AI?

We should create a "service" where an AI will represent us in the interview process and then we can just have an AI talk to the AI. We'll do our best, but if our AI just straight up starts lying to make us look good, well, what can you do?

Exactly what I was thinking. Have a deepfake of yourself, driven by an AI trained on your work history and the common interview questions, ready to take over your session when you spot an AI.