Artificially intelligent weapon systems will thus be capable of making their own decisions, for instance, about their target, or their approach to their target, and of doing so in an ‘intelligent’ fashion. While they will be programmed to make decisions according to certain rules, in important circumstances their actions will not be predictable. However, this is not to say that they will be random either. Mere randomness provides no support for a claim to autonomy. Instead the actions of these machines will be based on reasons, but these reasons will be responsive to the internal states —‘desires’, ‘ beliefs’ and ‘values’— of the system itself. Moreover, these systems will have significant capacity to form and revise these beliefs themselves. They will even have the ability to learn from experience.