Family Learning Talk in AI Literacy Learning Activities

Title of the Paper

Family Learning Talk in AI Literacy Learning Activities

Paper Information

• Subject Area: Informal Learning in Human-Computer Interaction and Artificial Intelligence Education
• Keywords: AI Literacy, Collaborative Learning, Interaction Design, Informal Learning, Family Learning, AI Education

Research Background and Issues

• Problem or Challenge:

  • Artificial Intelligence (AI) plays an increasingly significant role in decision-making in areas such as hiring, recommendation systems, law enforcement, and military applications. However, public understanding of AI systems remains insufficient.
  • Existing AI education efforts are largely focused on K-12 classrooms, with limited research on informal learning, particularly in family learning contexts.

• Significance:

  • AI literacy can help the public better evaluate AI technologies, engage in human-computer interaction, and voice their opinions on AI-related public issues.
  • Learning activities within families can foster intergenerational dialogue, aiding both children and adults in understanding the social and ethical implications of AI.

• Motivation and Related Work:

  • Informal learning spaces have the potential to attract a broader audience and provide an environment for collaborative family learning.
  • The authors build on their prior research on AI literacy education design principles to offer new perspectives for designing AI education activities in informal learning contexts.

Solution

• Methods or Solutions:

  • Designed three AI education activities (Knowledge Net, Creature Features, and LuminAI) to explore the types of dialogue within family groups in informal learning environments.
  • Each activity was designed based on specific AI literacy design principles, such as embodied learning, collaborative interaction, and transparency.

• Innovations:

  • Introduced full-body interaction and tangible learning tools (e.g., physical cards, weight markers) into AI education activities.
  • Proposed a new learning dialogue analysis framework to evaluate family learning conversations in informal learning environments.

• Implementation Steps and Key Technologies:

  1. Knowledge Net: Utilized a physical network representation method, where a chatbot answered family questions to simulate AI decision-making.
  2. Creature Features: Designed a training dataset to classify birds using a feature-based machine learning algorithm. Users modeled data using cards and weight markers.
  3. LuminAI: Enabled full-body interaction with an AI dance partner, allowing users to explore the decision logic and visualized memory of AI dance moves.
  4. Data collection and dialogue analysis were conducted using interview-based and video-recorded research methods.

Research Outcomes

• Specific Findings:

  • Analyzed family dialogue data from the three activities, revealing the effectiveness of different activities in conveying AI literacy skills (e.g., knowledge representation, decision logic).
  • Provided empirical support for design principles such as transparency, embodied interaction, and social interaction.

• Advantages Compared to Existing Solutions:

  • The Knowledge Net activity supported extended collaborative interaction among families with low technological literacy.
  • Creature Features demonstrated how families could discuss AI decision-making and the limitations of machine learning through iterative dataset adjustments.
  • LuminAI expanded the AI education experience through creative interaction, sparking discussions about whether AI can possess creativity.

• Experimental or Evaluation Results:

  • The study analyzed types of "learning talk" in participants' dialogue, including judgment, integration, and generation.
  • Identified which activities most effectively facilitated conceptual discussions and skill comprehension.

• Limitations and Future Directions:

  • Limitations included technical issues (e.g., video recording failures) and a small sample size.
  • Future research should explore how to support parents' understanding of technical language and their educational roles, as well as how to engage a broader audience in AI education activities.
  • Investigate how to balance embodied learning with digital interaction to optimize overall learning outcomes.

Overall Evaluation

This paper provides a detailed analysis of AI literacy education, presenting diverse AI education activity designs and principles, and thoroughly exploring family group learning in informal settings. The research findings offer significant insights for advancing the design of AI education tools and promoting public dissemination of AI knowledge.