Counterspeakers’ Perspectives: Unveiling Barriers and AI Needs in the Fight against Online Hate

Title of the Paper

Counterspeakers’ Perspectives: Unveiling Barriers and AI Needs in the Fight Against Online Hate

Paper Information

• Field: Human-Computer Interaction (HCI), Online Action Against Hate Speech, AI Support
• Keywords: Hate Speech, Counterspeech, AI-Supported Counterspeech, AI-Mediated Communication, Online Activism

Research Background and Issues

• Problems and Challenges: The scale and complexity of online hate speech continue to grow. Traditional content moderation mechanisms (e.g., removal or blocking) often raise concerns about censorship and fail to comprehensively address the issue of hate speech. Counterspeech, as an alternative approach, employs proactive responses and community interaction to avoid the negative impacts of censorship. However, its implementation faces significant challenges, such as limited resources, personal risks, and difficulties in evaluating effectiveness.
• Significance: As hate speech proliferates in online spaces, these issues have become critical for fostering a healthy online environment. The involvement of AI in supporting counterspeech and its potential risks warrant in-depth investigation.
• Research Motivation: Existing studies lack a systematic understanding of user needs and the role of AI in counterspeech practices. This research aims to explore barriers in counterspeech and how AI can assist in overcoming these challenges by engaging counterspeech activists and regular social media users.

Solution

• Methodology:
  • Semi-structured interviews: Conducted in-depth interviews with 10 long-term counterspeech activists to gain insights into their experiences.
  • Large-scale survey: Distributed questionnaires to 342 regular social media users to capture broader counterspeech participation experiences.
• Innovations:
  • First systematic study of user perspectives on counterspeech, including barriers, motivations, and needs for AI tools.
  • Proposed a three-step theory of counterspeech participation (deciding to respond -> constructing a response -> posting and handling follow-up).
  • Synthesized user expectations for AI assistance, suggesting design considerations such as enhancing user capabilities, emotional support, and fact-checking.
• Implementation Steps:
  • Data collection and analysis from interviews and surveys employed a mixed qualitative and quantitative approach, including open coding and grounded theory.
  • Identified functional characteristics required for AI tools in counterspeech based on user needs and feedback (e.g., information aggregation, emotional management, and automated response generation).

Research Findings

• Specific Findings:
  • Counterspeech barriers were categorized into four types: limited resources (e.g., time, energy), insufficient training (e.g., how to construct effective speech), unclear impact (e.g., difficulty in observing outcomes), and personal harm (e.g., psychological stress).
  • AI tools were envisioned as a way to alleviate these barriers:
    • Functional needs include guiding user decisions (e.g., identifying hate speech), assisting in constructing personalized counterspeech, supporting emotional communication, and verifying facts.
    • Significant differences in AI tool acceptance and preferences were observed between activists and regular users, with the latter favoring automated solutions.
  • Users expressed multiple concerns about AI involvement, primarily:
    • Lack of authenticity and trust: AI-generated content might be perceived as insincere or untrustworthy.
    • Absence of moral agency: Using AI could diminish users’ sense of moral responsibility.
    • Functional limitations: For example, AI’s constraints in cultural context and emotional intelligence.
• Advantages:
  • Enhanced the effectiveness and reach of counterspeech, encouraged public participation, and reduced psychological stress for users.
  • Unlike fully automated counterspeech generation, this study emphasizes the importance of human-AI collaborative design to ensure authenticity.
• Limitations and Future Directions:
  • Limitations: The study sample was primarily from North America, limiting generalizability to other cultures and regions; the broad definition of AI did not explore users’ varied understandings of the technology.
  • Future Directions:
    • Investigate the impact of cultural contexts on counterspeech and the adaptability of AI tools.
    • Develop AI systems with greater transparency, fairness, and cultural sensitivity, especially for minority groups.
    • Explore counterspeech strategies across different platforms and content formats (e.g., short videos).

This study reveals the complexity of online counterspeech, clarifies the potential and limitations of AI tools, and provides valuable insights for designing technologies that balance effectiveness, ethics, and user well-being in combating hate speech.