Designing Chatbots with Black Americans with Chronic Conditions: Overcoming Challenges against COVID-19

Document Title

Designing Chatbots with Black Americans with Chronic Conditions: Overcoming Challenges against COVID-19

Document Information

• Subject Areas: Human-Computer Interaction Design, Health Technology Design, Culturally Sensitive Technology Design
• Keywords: Black American community, chronic conditions, chatbots, COVID-19, participatory design

Research Background and Issues

• Identified Issues:

  • Chatbots are widely used in the health domain, but a lack of culturally tailored designs may result in low user trust and willingness to engage.
  • Designing chatbots for marginalized communities faces challenges due to insufficient understanding of cultural and socioeconomic contexts, limiting the potential of technology to reduce health disparities.

• Why It Matters:

  • The Black American community is particularly vulnerable to chronic conditions and the impacts of COVID-19. Culturally sensitive health technology design can improve access to and trust in health information, helping to alleviate health inequities.

• Research Motivation:

  • The authors aim to explore how participatory design can create more suitable chatbots for the Black American community, improving health management, especially in the context of COVID-19.

• Related Work:

  • Existing literature shows that understanding cultural contexts can significantly enhance the effectiveness of health technology design. However, most work focuses on mobile apps or electronic health record systems, with limited research on chatbots.
  • Studies on designing technology for Black communities emphasize participatory methods and collective community culture, but rarely address the design of AI chatbots.

Proposed Solution

• Approach to Address Issues:

  • Using a participatory design approach, the authors engaged Black Americans in chatbot design activities to identify their needs and expectations.
  • Semi-structured interviews and design activities were conducted to understand user expectations regarding chatbot roles, functionalities, and usability challenges.

• Innovations:

  • Developed a platform (botframe.com) to allow participants to design and simulate imagined chatbot conversations, providing deeper insights into their design thinking.
  • Proposed specific design recommendations, including transparency in information sources, community data integration, and consideration of socioeconomic factors to optimize chatbot design.

• Implementation Steps:

  1. Recruit eligible Black American participants (with chronic conditions and at high risk for COVID-19).
  2. Conduct research activities in three parts: pre-interaction activities (experiencing a chatbot via Facebook Messenger), semi-structured interviews, and chatbot design activities.
  3. Participants used the botframe platform to simulate and design chatbot conversations based on six defined scenarios.
  4. Data from interviews and activities were analyzed using open coding and inductive methods to summarize requirements and design challenges.

Research Findings

• Specific Findings:

  • Desired Roles: Chatbots as intermediaries connecting health services, personalized health advice assistants, information hubs, trusted community voices, and mental health management tools.
  • Key Functional Requirements:
    1. Having a trustworthy and comfortable "personality."
    2. Remembering user interactions and adapting future responses.
    3. Communicating in a direct and unbiased manner.
    4. Collecting user data through questions and helping users reflect.
  • Potential Barriers to Use: Difficulty in building trust, over-reliance on chatbots, inadequate community infrastructure (e.g., network bandwidth issues), and personal health challenges (e.g., vision impairment).

• Advantages Compared to Existing Solutions:

  • Provided more specific culturally sensitive design guidelines, such as integrating community data and enhancing transparency to address trust issues.
  • Highlighted the importance of chatbots serving not only individuals but also addressing community-wide needs, a perspective rarely seen in existing technology designs.

• Experimental or Evaluation Results:

  • Analysis of participants' imagined chatbot designs revealed practical functionalities and design suggestions suitable for the Black community.
  • Participants emphasized challenges and solutions related to integrating community needs.

• Limitations and Future Directions:

  • Limitations: Participants were primarily Black women from the Midwest, limiting generalizability to the broader Black American population. The study also did not include participants from low-income or less-educated backgrounds.
  • Future Directions:
    1. Expand research to include low-income and lower-education user groups, particularly those who rely solely on smartphones.
    2. Explore new technological solutions, such as offline data interaction modes, to increase accessibility.

Design Recommendations

1. Transparency of Information Sources: Clearly display where the chatbot retrieves information (e.g., CDC, social media, or community contributors) to build user trust.
2. Community Data Integration: Chatbots should record and analyze collective community needs to design solutions targeting the community as a whole.
3. Consideration of Socioeconomic Factors:
   • Provide multiple health options based on users' economic conditions.
   • Support low-bandwidth or offline application scenarios.
   • Offer mental health assistance and guide users' thought processes through system Q&A.

Conclusion

This study combines culturally sensitive participatory design methods to propose chatbot design recommendations tailored to the Black American community, particularly addressing the needs for health service accessibility, cultural trust, and socioeconomic barriers in the context of COVID-19. This work highlights the potential of AI technologies to serve marginalized groups and advances the HCI community's efforts in promoting diversity in design.