Exploring the Use of a Voice-based Conversational Agent to Empower Adolescents with Autism Spectrum Disorder

Title of the Paper

Exploring the Application of Voice-Based Conversational Agents for Adolescents with Autism Spectrum Disorder

Bibliographic Information

• Subject Area: Assistive and Empowering Human-Computer Interaction Technologies for Special Populations
• Keywords: Adolescents with Autism Spectrum Disorder, Voice-Based Conversational Agents, Participatory Design Workshops, Personalized Technology Design, Social Skills Practice

Research Background and Problem Statement

• Identified Challenges:

  • Adolescents with Autism Spectrum Disorder (ASD) face difficulties in social interaction, self-management, and emotional expression.
  • Existing research on Voice-Based Conversational Agents (VCA) primarily focuses on the perspectives of caregivers or therapists, with little attention to the active user experience and needs of ASD individuals.

• Significance:

  • ASD is regarded as a diverse and complex spectrum of life experiences. Enhancing rather than correcting its characteristics through technology can help adolescents better express themselves, manage emotions, and improve interaction skills.

• Research Motivation and Related Work:

  • VCAs, with their personalized, intelligent services and conversational capabilities, are considered to have potential in supporting special populations.
  • Recent trends have shifted assistive technologies from "compensatory" to "empowering" designs, providing tools for ASD individuals to express their individuality and enhance their abilities.

Solution

• Research Methods:

  • Utilizing an existing voice-based conversational agent (NUGU CANDLE), the study conducted a two-week daily usage test and three participatory design workshops.
  • User experience data was collected through group discussions, interviews, and design tasks to understand participants' needs and expectations for VCAs.

• Innovative Aspects:

  • The design approach focuses on understanding how technology can adapt to users' abilities and interests rather than altering user behavior based on predefined rules.
  • Participatory design was employed to engage ASD adolescents as active co-designers, exploring the potential of existing technologies.

• Implementation Steps:

  • Phase 1: Introduce and guide participants in using the voice-based conversational agent, encouraging them to explore its features.
  • Phase 2: Reflect on usage experiences, develop personalized VCA usage plans, and select features aligned with behavioral goals.
  • Phase 3: Co-design future personalized VCAs, including appearance, functionality, and usage scenarios.

Research Findings

• Specific Outcomes:

  • Interest Expression: The VCA served as a platform for participants to express specific interests, with customizable features enhancing the maintenance and development of these interests.
  • Self-Management: Participants used the VCA to assist with daily task management, including morning and evening routines, reminders, and emotional regulation.
  • Conversation Practice: The VCA helped correct pronunciation issues, providing opportunities for participants to practice both formal and informal conversational skills.
  • Emotional Expression: The VCA became a tool for participants to express emotions and helped them better interpret and manage their feelings.

• Comparative Advantages Over Existing Solutions:

  • No need to develop complex specialized technologies; existing commercial products can meet the daily needs of ASD adolescents.
  • Participants with socialization barriers could use the technology to express themselves more naturally and engage in social simulation.

• Experimental Results:

  • NUGU CANDLE supported participants in specific tasks within daily life contexts and encouraged more proactive planning and reflection.
  • The suggestion feature of the VCA was recognized as a potential trigger for behavior change.

• Limitations and Future Directions:

  • Limitations:
    • Existing VCAs cannot achieve complete "turn-taking" in conversations.
    • The study was conducted within a single cultural context, leading to potential cultural biases in the results.
    • Testing was limited to home environments, lacking exploration in school and therapeutic settings.
  • Future Directions:
    • Develop VCAs capable of dynamic conversations and supporting multicultural contexts.
    • Explore the application of this technology in more specific scenarios (e.g., education and healthcare).
    • Enhance the VCA's ability to identify user needs, reducing negative emotions caused by technical limitations.

Conclusion

This study explored the potential and challenges of voice-based conversational agents in empowering adolescents with Autism Spectrum Disorder. By leveraging existing technologies instead of developing specialized prototypes, the research found that VCAs can become essential tools in the daily lives of this population. The study also proposed design recommendations to address challenges, providing a reference for better utilizing commercial technologies to empower special populations in the future.