Arrow’s Impossibility Theorem often gets brought up in such discussions, but less commonly mentioned is the fact that it only applies to ranked voting systems and not score-based / cardinal voting systems. It also says that a voting system can't produce a total ranking of all options (which is a stronger requirement than just picking a single winner), and it assumes "independence of irrelevant alternatives"[0] which seems rational but doesn't reflect how humans actually behave.
https://en.wikipedia.org/wiki/Independence_of_irrelevant_alt...