A Perspective on Building Ethical Datasets for Children's Conversational Agents.

Area of Science:

• Developmental Psychology
• Human-Computer Interaction
• Artificial Intelligence Ethics

Background:

• Artificial intelligence (AI) is increasingly integrated into children's environments, influencing social and learning experiences.
• Children frequently interact with AI through conversational agents, necessitating socially intelligent designs tailored for young users.
• The data used to train these AI systems for children often lack transparency.

Purpose of the Study:

• To apply a sociocultural developmental approach to analyze child-centric intelligent conversational agents.
• To explore how children's developmental stages impact their social learning and interaction with AI.
• To discuss data types for training AI speech generation and ethical considerations.

Main Methods:

• Sociocultural developmental framework application.
• Analysis of children's developmental influences on AI interaction.
• Examination of potential data types for AI training and ethical implications.

Main Results:

• Children's development significantly shapes their social learning and AI engagement.
• Specific data types can be curated to train AI for child-appropriate conversational agents.
• Ethical considerations are paramount in dataset curation and AI model training for children.

Conclusions:

• Developing socially intelligent AI for children requires understanding their developmental needs.
• Ethical data practices are crucial for responsible AI development in children's technology.
• Future research should focus on advancing social AI for enhanced child development and learning.