Fairness by Design: Cross-Cultural Perspectives from Children on AI and Fair Data Processing in their Education Futures

Research Background and Issues

• What problems or challenges did the authors identify?
  AI-driven educational technology (AI-EdTech) is widely applied in schools for personalized learning, real-time attention monitoring, and progress tracking. However, it poses risks such as data privacy concerns, commercial exploitation of children's data, and threats to children's autonomy and mental health.
  The fairness principle in current data protection laws lacks clear definitions, making it difficult for designers to translate fairness into concrete design practices.
  Greater attention is needed to understand how children's own perceptions and expectations of fairness can be reflected in AI-EdTech design, particularly considering perspectives from cross-cultural groups of children.

• Why is this issue important?
  Education is a critical domain for children's development, and the widespread use of AI-EdTech could significantly impact their privacy, rights, and future opportunities. If misused or poorly designed, these technologies may lead to data bias and discrimination, exacerbating power imbalances in education. Moreover, children's voices have not been adequately incorporated into design and legal practices.

• Research Motivation and Related Work
  Existing studies primarily focus on algorithmic bias or data privacy, while children's understanding of fairness remains underexplored. Using participatory design methods, investigating children's perceptions of fairness from a cross-cultural perspective can provide valuable insights for technological design and legal regulation.

Solution

• What methods or solutions did the authors propose?
  Through cross-cultural participatory design workshops, the authors explored the perspectives of children aged 10 to 12 on fair data processing and their expectations for AI-EdTech systems. These workshops included design activities where children imagined future "fair" AI-EdTech systems and proposed related solutions.

• What is innovative about this solution?

  • Defining fairness from children's perspectives rather than relying on adult viewpoints or legal frameworks.
  • Conducting cross-cultural comparative studies to examine how national and cultural contexts influence perceptions of fairness.
  • Employing future-oriented design methods (Participatory Design Futuring) to help children envision AI systems beyond current technological limitations.
  • Integrating design principles such as transparency, data control, and personalization to better align systems with children's needs.

• What are the implementation steps and key technologies used?

  • Designing three workshop activities, ranging from basic concepts (e.g., personal data and privacy) to imagining future AI designs.
  • Collecting data through children's discussion recordings, drawings, letters, and researchers' observational notes.
  • Conducting workshops across schools and countries (Scotland and Turkey) to analyze children's views on fair data processing.
  • Using reflective and thematic analysis methods to synthesize insights, presenting children's perspectives through visual tools and direct product designs.

Research Outcomes

• What specific outcomes were achieved?

  • Identified children's expectations for data transparency, personal data control, and personalized support. They emphasized that AI-EdTech fairness should address emotional well-being, transparent data usage, and accountability.
  • Proposed concrete solutions, such as the "Data Hair Robot," which allows children to intuitively understand the amount of data collected and control its usage.

• How does it compare to existing solutions?

  • Focuses on children's perspectives, offering a unique and detailed definition of fairness.
  • Highlights interactive and multimodal design (e.g., music, visual representation) to enhance children's engagement and comprehension.
  • Expands the concept of fairness beyond preventing data harm to actively empowering children.

• What were the experimental or evaluation results?

  • Collected extensive data from 76 participants in Scotland and Turkey, revealing that children's expectations for transparency, diversity, and personalization were consistent across cultures, while specific considerations varied based on individual values.
  • Cross-cultural findings illustrated how children perceive the fairness of data exchange, such as whether data is used for public or personal benefit.

• Limitations and Future Directions

  • Limitations: The study only examined samples from two countries, lacking broader cultural representation; the participant pool and age range were limited, making it difficult to fully represent global children.
  • Future Directions: Expand the age range and cultural scope; design activities less influenced by researcher biases to explore children's concepts of fairness across diverse socioeconomic backgrounds.

Conclusion

This study, through participatory design, revealed children's unique perspectives on the fairness of AI-EdTech data practices. It proposed innovative design methods and legal policy recommendations, offering valuable references for child-centered design and data protection. Future research should further explore children's understanding of fairness across cultural, economic, and regulatory dimensions, and develop more universally applicable fairness design frameworks.