ABSTRACT
With the recent, widespread release of generative Artificial Intelligence (AI), people of all ages have sought to make sense out of interactions with computer programs that now approximate human quality communication. As children interact with a non-living entity that sounds reasonably human, it is reasonable to question whether children engage in similar social reasoning with AI as they do with people in their lives. Social Domain Theory identifies three domains of reasoning, moral, social conventional, and personal, that involve distinct concepts in coordinating social judgements. The personal domain, and in particular the expansion of the personal domain during adolescence, is often used in choosing to whom one discloses sensitive information. Learning how children conceptualise AI and reason about personal disclosure to AI bots is of great importance in providing for safe child/AI interactions and informing the choices of policymakers and those responsible for developing AI safeguards. Preliminary analyses of survey data from American elementary (8–10-year-olds) and high school (16–18-year-olds) children are presented, showing differences in knowledge and conceptualization of AI by age group. Differences in children’s tendency to engage in social reasoning with AI as well as implications for personal disclosure are discussed.
