An AI, which would be hard to imbue with relevant strategic and political knowledge, might not be able to judge for itself whether an attack was permitted.

人工智能很难被灌输相关的战略和政治知识，它可能无法自己判断攻击是否被允许。