Well at least to some people, this makes it look like a sleazy attempt form customer service at deflecting blame from a fact ("the steak is undercooked") to a feeling from the customer ("you just don't like the steak, but I don't believe you when you say it's undercooked").
It immediately makes the person seem less human and more like a customer service robot. I'm pretty sure most people hate it, but maybe I'm wrong.
If you assume I can take a good look at you and just know you're the kind of guy who would never do that, you're assuming a level of sight-reading people that even most police investigators don't have. I'm sorry, I'm only human, and I'm waiting five tables simultaneously right now.
If the steak is blue and they ordered medium ... then there is little room for debate. If they wanted something other than what they ordered, then validating the feelings is more appropriate.
And that debate can be had (or not) by a lawyer or perhaps a manager, whose job it is to do so. No server is going to be vested with that authority, nor wants to be saddled with the uncompensated responsibility to.
Which can be a mistake when the person you are dealing with has or may have an ulterior motive for your interaction (i.e. said "toddlers").
This is why in actual customer service, validating someone's feelings ("I understand you did not like the cook on the steak") is good, while validating their concerns ("I understand that the steak was undercooked") is bad.
You don't want to "find common ground" or "shared viewpoints" just to fulfill the validation matrix plot, because it may very well be based on a false premise, or even a blatant fabrication. In real world terms, validating concerns can often be an admission of liability or fault, or a soundbite that will be weaponized against you.