C3AI 2nd Workshop

What human factors shape children's subjective feelings of trust and distrust towards AI? How can we create trustworthy AI systems that not only foster positive relationships but also help children understand AI limitations and manage their expectations? How do children decide when to trust, doubt, or challenge an AI system? What developmental factors shape the little moments where kids show confidence, curiosity, hesitation, or critique? 

The second edition of the C3AI workshop explores how trust and distrust emerge in real child–AI conversations. We look closely at the human factors behind these interactions: cognitive differences across ages, emotional reactions such as confusion or confidence, social expectations, and the lived experiences that shape how children interpret AI behaviour. Even though today’s conversational AI systems aren’t designed for children, kids use them anyway, at home, in school, and in creative or social activities. This makes it essential to build evaluation approaches that are transparent, ethical, and developmentally sensitive.

In this workshop, participants will work hands-on with authentic child–AI transcripts, age-diverse personas, and real-world scenarios to analyse how trust unfolds over time. Together, we’ll co-design early versions of evaluation metrics that help us understand:

• how children across different ages interpret AI actions,

• how trust is built, lost, or repaired,

• how children express uncertainty or doubt,

• how systems can better support reflection, safety, and agency.
  
  

Our shared goal is to move toward child-centred evaluation frameworks that respect developmental diversity, uphold children’s rights, and support responsible, transparent AI design. Participants will help build reusable resources that empower researchers, designers, and educators to create AI systems that children can navigate confidently and critically.