The Hall of AI Fears and Hopes: Comparing the Views of AI Influencers and those of Members of the U.S. Public Through an Interactive Platform

Research Background and Issues

• What problems or challenges did the authors identify?

  • The development of artificial intelligence (AI) is dominated by academic and industry leaders (referred to as "influencers"), but it remains unclear whether their perspectives align with public opinions on AI.
  • Notably, there may be discrepancies in views among different social groups (e.g., women, people of color), exacerbating the issue of unrepresentative decision-making in technology.

• Why is this issue important?

  • The public is directly impacted by the implementation of AI technologies, yet their involvement in AI development and governance is often limited or entirely absent. A lack of public participation may result in ineffective policies or public rejection, ultimately undermining societal trust in AI.
  • The design and governance of AI systems need to be more inclusive to avoid inequalities stemming from neglecting public concerns.

• Research Motivation and Related Work

  • Existing research primarily focuses on analyzing the perspectives of either the public or AI practitioners individually, with limited studies comparing the views of the public and AI influencers.
  • This study aims to fill this research gap by exploring whether the perspectives of the public align with those of key decision-makers.

Solutions

• What methods or solutions did the authors propose?

  • The authors developed an interactive data collection platform called "The Hall of AI Fears and Hopes" to gather public opinions on AI hopes and fears and compare them with the perspectives of AI influencers.

• What is innovative about this solution?

  • Instead of using traditional surveys, the platform employs interactive visual design to guide participants in exploring AI influencers' perspectives and reflecting on their own views.
  • It establishes metrics to analyze differences in perspectives between the public and various influencer groups (e.g., younger vs. older influencers, billionaires vs. non-billionaires).

• What are the implementation steps and key technologies used?

  1. Data Preparation: Select 100 AI influencers identified by Time Magazine as the research sample and annotate their hopes and fears regarding AI.
  2. Platform Design: Develop a six-part interactive platform through literature review and multiple rounds of co-design, including modules for customizing AI definitions, self-assessment of AI opinions, and matching participant views with influencer perspectives.
  3. Experiment Design: Conduct a large-scale online experiment via Prolific, recruiting 330 participants representative of U.S. demographics. Collect their opinions through the platform and analyze the alignment with AI influencers' perspectives.
  4. Method Evaluation: Analyze public and influencer perspectives from both qualitative and quantitative angles and define alignment metrics to calculate discrepancies between the two groups.

Research Outcomes

• What specific results were achieved?

  • The public's primary fears about AI include loss of control, increased unemployment, and misinformation, while AI influencers focus more on the necessity of regulation and the potential for technological innovation.
  • Alignment analysis revealed that younger influencers' perspectives align most closely with the public, whereas older influencers and billionaires show the lowest alignment.
  • Despite increased representation of marginalized groups (e.g., women, people of color) among AI influencers, significant gaps remain between their views and those of the public with similar backgrounds.

• What advantages does this solution have compared to existing ones?

  • The platform is more engaging and interactive, significantly enhancing participant interest, particularly in helping the public form a more comprehensive understanding of AI compared to traditional surveys.
  • It introduces a novel method for quantifying the alignment between public and influencer perspectives, providing a clear analytical framework for future research.

• What were the experimental or evaluation results?

  • The platform received high satisfaction ratings, particularly for balancing perspective expression and engaging participants across multiple levels (rating range: 4.94/7 - 5.66/7).
  • Public opinions primarily focus on AI's impact on employment, privacy, and social services, while influencers exhibit a broader systemic perspective, such as global governance and innovation benefits.

• Limitations and Future Directions

  • The platform's applicability may be limited, as the study is based solely on U.S. population samples; future research should expand to international groups.
  • The AI influencers selected by Time Magazine are restricted to publicly available data, potentially failing to represent the diversity of global AI perspectives.
  • Future work could further optimize tool design to address potential complexity and accessibility issues while exploring more efficient methods for gathering public opinions, such as natural language processing techniques based on corpora.

This study not only highlights the potential disconnect between public and key decision-makers' perspectives but also provides innovative methods for designing AI governance and enhancing public participation.