With the advance of deepfakes, we can now relatively easily swap faces of one person to the other. I genuinely believe that looks are everything when it comes to perception and bias. I also think that race isn't necessarily as much of a factor in the US (though there may definitely be some).

Why don't we put these to test, and even try to quantify it. Take a speech of any person, and swap its faces and try to control for what looks do affect what perception the most. Those experiments could be used for dating biases, as well as anything else such as "muh confidence", desire to get to know/befriend that person, leadership biases, or even competence judgements (and many other things probably).

Before the results are in, what do you guys think will come out? Does look really not matter that much? Meaning swapping faces won't impact group perception too much? Or does it really drive most of people's judgments.