But then, to see what the allow-no-harm rule could accomplish in the face of a moral dilemma, he presented the A-robot with two H-robots wandering into danger simultaneously. Now how would it behave? ... In almost half of the trials, the A-robot went into a helpless dither and let both 'humans' perish.
So the A-robot is vulnerable to the paradox of Buridan's ass ("an ass which, confronted by both food and water must necessarily die of both hunger and thirst while pondering a decision"). In these cases, it can help to add more factors that make one option clearly better (is one human a baby? is one of them suicidal? is one dying of cancer anyway?), but if determining these factors is time-consuming and impractical, it would be better to just flip a coin.