At work I already had teammates who would grant LGTM to code they barely read, but looked fine after reading the description and hopefully also skimming over the code. I feel that this path of least resistance/effort will also apply to things that an LLM spit as those are highly likely to look correct on the surface level.