I do not quite understand the question.
Are you referring to the general problem to have the same private key generated twice by "accident"? This is so extremely unlikely that we could exclude it. See
https://bitcointalk.org/index.php?topic=233503.0Or are you referring to the "provability" that we did not create the same private key before somewhere else. This is easy to answer: The key generation is taking place outsite of Bitalo in the user's browser and encrypted there. As the code is in javascript and executed in the browser (= open source and publicly verifiable) it can be proved that the user private key is NEW and random.
You can follow the single steps by decoding the executed browser commands.