Exploring Effects of Chatbot-based Social Contact on Reducing Mental Illness Stigma

Title of the Paper

Exploring Effects of Chatbot-based Social Contact on Reducing Mental Illness Stigma

Paper Information

• Research Domain: Using chatbots to promote social contact and reduce mental illness stigma
• Keywords: Chatbot, social stigma, mental illness, self-disclosure, storytelling

Research Background and Problem Statement

• Identified Issues or Challenges:

  1. Stigma surrounding mental illness (e.g., perceiving patients as dangerous or responsible for their condition) hinders help-seeking and recovery.
  2. Existing anti-stigma interventions, such as public contact activities, are often limited by resources and scalability.
  3. While chatbots have shown potential in mental health interventions, their effectiveness in reducing social stigma remains underexplored.
  4. It is unclear how first-person versus third-person chatbot interactions impact stigma reduction.

• Importance and Research Motivation: Reducing stigma related to mental health can improve patients' social support networks and help-seeking behaviors. Cost-effective and scalable chatbots could serve as innovative tools for stigma reduction. Exploring their design optimization offers new perspectives for mental health interaction research.

• Related Work:

  • Narrative approaches (e.g., storytelling) and social contact have proven effective in reducing mental illness stigma.
  • Chatbots have been utilized in areas such as cognitive behavioral therapy and anonymous mental health support, but direct interventions targeting stigma are scarce.

Proposed Solution

• Proposed Solution: The authors designed two types of chatbots: one interacting in the first person as a character facing mental health challenges, and the other narrating these stories in the third person as an information mediator.

• Innovative Contributions:

  1. Introducing and comparing two different narrative roles for chatbots (first-person and third-person) in influencing social stigma.
  2. Simulating social contact through chatbots to test their potential in promoting self-disclosure and reducing stigmatizing attitudes.
  3. Employing a comprehensive evaluation approach combining survey data, interaction logs, and interview findings.

• Implementation Steps:

  1. Story Design: Seven mental health stories about the character Kenta were developed based on scenarios and symptoms outlined in the DSM-5 and previous research.
  2. Study Groups: Participants were divided into three groups: those using the first-person chatbot (FP), those using the third-person chatbot (TP), and a control group using a web-based survey.
  3. Data Collection: A two-week experiment was conducted, including daily task completion logs, survey responses, and follow-up interviews.
  4. Evaluation Metrics:
     • Whether the intervention facilitated self-disclosure.
     • Changes in participants' stigmatizing attitudes, including attribution of responsibility, emotional responses (e.g., empathy, fear), and behavioral intentions (e.g., willingness to help).

Research Findings

• Specific Findings:

  1. The first-person (FP) chatbot was more effective in fostering social contact, eliciting greater self-disclosure from participants, and promoting empathy and support for the story's character.
  2. Compared to the control group, both FP and TP chatbots significantly reduced participants' attribution of responsibility for mental illness (shifting from "internal attribution" to "external attribution").
  3. The FP chatbot significantly decreased participants' fear of patients and reduced feelings of social distance. In contrast, the TP chatbot contributed to increased willingness to help but was less effective overall than the FP chatbot.
  4. The web-based survey unexpectedly increased social distance among some participants toward individuals with mental illness.

• Advantages:

  • The first-person chatbot better simulates "social contact," guiding users to perceive the character as a "trustworthy entity," thereby reducing stigmatizing attitudes.
  • Chatbots' cost-effectiveness and scalability offer technological potential for large-scale anti-stigma interventions.

• Experimental and Evaluation Results:

  • Data from three sources—pre/post-test survey responses, interaction logs, and semi-structured interview results—confirmed the FP chatbot's advantages in reducing stigma, fostering empathy, and encouraging self-disclosure.

• Limitations and Future Directions:

  1. The use of a single character (Kenta) may limit coverage of diverse mental illness experiences.
  2. Some users may experience emotional distress from severely negative scenarios (e.g., suicidal thoughts), potentially impacting intervention effectiveness.
  3. Chatbot role settings (e.g., first-person, third-person) need optimization based on specific contexts. Future research could explore broader applications, including targeting other stigmatized groups (e.g., LGBTQ communities, individuals with physical disabilities).