Yeah, one of the things that has become apparent about these low-effort AI responses is the tendency for the output to contain a lot of "may"/"could"/etc terminology. AI is good for summarizing without much effort, but thinking it's a slam-dunk debate winner is laughable.