The benchmark used by the Office for Nuclear Regulation for judging how much should be spent on nuclear safety has no basis in evidence and places insufficient value on human life, suggests a new paper. The review suggests it may need to be 10 times higher — between £16 million and £22 million per life saved.