I use AES quite often at work (I'm a software programmer) and I trust that it "works" without understanding the maths behind it. It's a black box to me.
Does a mathematical proof exist that AES 256 encryption works for any data?
I.e. Is it guaranteed that there is no "weird pattern of plain text" (input data) which when encrypted and then decrypted yields the wrong result?