Childhood and adolescence is a critical time for young people's development, as they learn how to form healthy relationships and develop critical thinking and self-regulation skills. AI chatbots, designed to foster intimate relationships through personalization and persuasion techniques, create issues to which children and young people are more likely to be vulnerable. This convergence of factors, related to both human development and technological design, heightens the possibility of young people's exposure to harm, including exposure to misinformation, sexualized and self-harm content, overreliance and deskilling, and the commercial exploitation of their data and privacy. As such, robust safeguarding solutions, designed specifically for children and young people, are needed in this space.
Related themes:
Child and Family Mental Health Access and Support
Children’s Well-Being and Safety in Modern Environments
Research suggests children and young people interact with AI chatbots differently from adults, and specific design features of these chatbots heighten their risk of encountering harm.
Current US policy does not specifically address the needs and rights of children and young people in relation to AI chatbots. Instead, there is a created expectation by AI companies that parents manage their children's interactions with AI chatbots.
As a result, this study draws on two large-scale nationally representative surveys of U.S. mothers (one of 2,290 and one of 2,818), to gain understanding of their experiences in relation to their children's AI chatbot use and their recommendations for practical and substantial safeguards.
Mothers reported that they have significant concerns about the design of platforms increasing the risk of harm to children, that AI is frequently introduced outside the home, and that they do not have visibility of their children's activities relating to AI chatbots, making it difficult for them to act as mediators.
Among mothers whose children use AI for more than entertainment, a majority report concern in three developmental areas: problem-solving and independent thinking; social skills; and emotional regulation.
Mothers place less trust in AI companies to act in the best interests of children and families than in any other institution, and that distrust is consistent across the political spectrum, from very conservative to liberal mothers alike.
A majority of mothers would support a law requiring stronger safeguards for children using AI.
The responsibility for creating safe spaces for children and young people must be a shared one between lawmakers, AI chatbot companies, and families.
Recommendations from mothers include the need for age-appropriate design standards; clear disclosure and transparency; meaningful consent procedures, particularly in school settings; limits on data collection; restrictions on manipulative and engagement-driven design; and finally, a recognition that AI platforms have an ongoing duty of care to their users.
AI chatbots present unique risks to children and adolescents, making age-appropriate safeguards essential to protect their development, privacy, wellbeing, and safety.
Explore Count on Mothers reports — rapid polls and in-depth national studies.
Methodology Count on Mothers conducts nationally representative research with U.S. mothers, weighted to reflect the population and reported in aggregate. Research is led by a PhD + MPH team. Findings have informed policy, industry, and media, and entered the Congressional Record on childcare, paid leave, and technology policy.
