Sure, it can turn a blurry image into one that "looks" better. But like with all things AI, how do you know that this interpretation is accurate? In other wor

Sure, it can turn a blurry image into one that "looks" better. But like with all things AI, how do you know that this interpretation is accurate? In other words, prove that this process wasn't done in an arbitrary or probabilistic manner. The fact that details were added doesn't make them necessarily correct and accurate.