Abstract: Provided are analyzer (120) that analyzes person information in relation to a person included in an image having been captured by imaging unit (110), database (130) that stores in itself the person information and environment information indicating an environment in which the relevant information processing system is installed, utterance unit (140) that makes an utterance having an utterance content in response to the person information and the environment information, and reinforcement learning unit (150) that, for each combinations of the person information and the environment information, performs learning to update a first score corresponding to the utterance content, based on result information indicating a result of the utterance having been made by utterance unit (140), and utterance unit (140) makes the utterance having the utterance content associated with the first score having a maximum value of the first score for the combination.