Designing Conversational Agents: A Self-Determination Theory Approach

Title of the Paper

Designing Conversational Agents: An Approach Based on Self-Determination Theory

Paper Information

• Research Domain: Human-Computer Interaction Design, Conversational Agents, User Experience
• Keywords: Conversational Agents, User Experience, Self-Determination Theory, Design Guidelines, Psychological Needs

Research Background and Issues

• Issues and Challenges: In designing conversational agents (CAs), how can we address users' psychological needs (competence, autonomy, relatedness) to provide a positive user experience?
• Significance: CAs are rapidly integrating into various aspects of human life (e.g., home, education, healthcare), offering convenience while potentially causing negative technological, social, and psychological impacts. Understanding how to design CAs to foster positive experiences has become increasingly urgent.
• Motivation and Related Work: Existing research highlights the significant impact of meeting psychological needs on user experience and well-being. However, current design recommendations often lack specificity and practical applicability. Moreover, the academic community has yet to propose specialized design guidelines tailored to the needs of CAs.

Solution

• Research Methodology:

  • Phase One: Conduct interview studies based on Self-Determination Theory (SDT) to explore how users experience competence, autonomy, and relatedness.
  • Phase Two: Extract design guidelines from the findings to assist in designing CAs that support psychological needs.

• Innovations and Implementation Steps:

  1. Apply SDT to emphasize the importance of psychological needs in the design process, aiming to achieve designs that go beyond mere functional satisfaction.
  2. Translate findings into 10 actionable and practical design guidelines, grounded in user research data and rigorous analysis to enhance usability.

Research Outcomes

• Specific Findings:

  • Competence: Users' sense of competence is influenced by their understanding of the system's capabilities and the communication efficiency of the CA. Recommendations include designing features to educate users about CA capabilities.
  • Autonomy: Users' sense of autonomy is reflected in their control over interactions with the CA and data management, as well as the degree of personalized experiences. Suggestions include providing more customization options and opportunities for data management.
  • Relatedness: Although users currently have low expectations for CAs to support social functions, findings indicate that CAs can subtly enhance relatedness in contexts such as family entertainment and interaction.

• Design Guidelines:

  • Core recommendations include: providing personalized feature overviews, introducing new functionalities during conversations, offering clear capability demonstrations, designing seamless interactions, and supporting data management (see Table 2 in the paper for details).

• Advantages:

  • The guidelines offer specific and actionable design recommendations, addressing gaps in existing HCI and AI design guidance.
  • The study provides theoretical support for meeting psychological needs, paving the way for future research and practice in CA design.

• Experimental or Evaluation Results:

  • The study validates the connection between user needs and design recommendations through interview practices. Design suggestions related to competence and autonomy were prioritized, while relatedness was excluded from practical guidelines due to current technological limitations.

• Limitations and Future Directions:

  • The guidelines have not yet been evaluated by design experts, and their practicality requires further validation.
  • Future research could incorporate additional theories or integrate literature to provide broader theoretical support for CA design.
  • Further exploration is needed to apply the presented guidelines in actual design processes.

Conclusion

• Key Contributions:

  1. Proposed a CA design direction based on psychological needs, emphasizing the importance of competence, autonomy, and relatedness.
  2. Developed 10 SDT-driven design guidelines to complement existing UI and AI design standards, supporting CA design practices.

• Impact and Applications: This study offers actionable guidance for designers, advancing theory-driven practices in conversational agent design. It also provides a new perspective on the relationship between technology and psychological needs.