Randomised is not what makes it ethical. It's what makes the AB test statistically valid such that you can test your hypothesis.

Ethical would be to run the test, gauge the click through rate to checkout to find a winner, then for all those you charged the higher price to give them a discount code such that everyone gets the product at the same price. What you're testing is the click through rate of a price advertised further up the funnel.

Maybe later you can say ok the rates didn't drop off people seem ok with the higher price. Higher price for all. Or not as the AB test would reveal.