Go back and read about the malfeasance of the petitioners there. They do things like cite a study, neglecting later data from the same study where the effect they want no longer shows up in the better data.

Policy has to go beyond what is rigorously demonstrated. It's not like criminal law. Technologies are not innocent until proven guilty. Just because LNT cannot be demonstrated at the doses relevant to nuclear accidents doesn't mean regulators should assume no effect.