Like, for example, what if someone doesn’t like cheese on their sandwich? Or loves pineapple on ther pizza? (its honestly good when combined with ham)
That plus your height and body shape and gender (as they can hear from your voice) is gonna narrow down the possibilities.
I think cameras can be trained to detect things, especially now with better hardware.
However, this probably won’t be done for 99% people. At least to me, it sounds extremely niche and to be used for spying and stuff against suspects.