I wonder how much this is affected by culture ("a lot" would be interesting and "not at all" would be very interesting). It feels like it's a low-level innate communication thing (e.g. dogs do this too when communicating with humans, while inter-dog eye contact has different semantics).

I do notice that I look at the speaker in a zoom call, even when I have my own video off. I also notice in big calls where many people have their cameras off, it's common that a questioner or commenter will enable the camera before speaking, and then disable it again when their turn is over.

(I searched for "culture" and "language" before posting)