Exploring What People Need to Know to be AI Literate: Tailoring for a Diversity of AI Roles and Responsibilities

Research Background and Issues

• What problems or challenges did the authors identify?
  Current AI literacy research primarily focuses on the knowledge required by users and developers, while other roles (e.g., executive managers, policymakers, legal professionals) face increasingly complex responsibilities due to the introduction of AI. These roles encounter numerous unaddressed knowledge gaps in decision-making, regulation, and AI application processes.

• Why is this issue important?
  The rapid development of AI technology impacts various roles across different sectors of society. A lack of AI literacy can lead to poor decision-making, legal risks, and negative social or personal consequences. For example, a lawyer's failure to understand that ChatGPT-generated content may contain false information could result in misuse and harm their professional development.

• Research Motivation and Related Work
  AI literacy is critical for promoting Responsible AI, helping individuals and organizations maximize AI's benefits while minimizing its risks. However, current AI literacy research lacks in-depth focus on non-user and non-developer roles (e.g., managers, policymakers, legal teams). By broadening the perspective, the authors aim to define the knowledge required by these roles and highlight the gaps in existing AI literacy research.

Solutions

• What methods or solutions did the authors propose?
  The authors adopted a service design approach to create value flow models that map the roles and responsibilities involved in AI applications and analyze how AI complicates these responsibilities.

• What is innovative about this solution?

  • The approach starts from the responsibilities of roles rather than focusing solely on users or developers to explore AI literacy needs.
  • By using a service design method, the study identified 16 roles across 41 AI applications and pinpointed gaps in AI knowledge and skills related to increasingly complex responsibilities.
  • A role-centered AI literacy framework was proposed and matched against the content of existing literacy research.

• What are the implementation steps and key techniques used?

  1. Creating Value Flow Models: Analyze value flows in different AI systems to identify relevant roles and their responsibilities.
  2. Role Integration: Consolidate roles from various models to cover all AI applications and summarize responsibilities for each role.
  3. Mapping Responsibilities to Literacy: Collect AI literacy research from the past nine years, match responsibilities with defined key AI literacy competencies, and identify areas of insufficient coverage.

Research Outcomes

• What specific outcomes were achieved?

  • Identified key roles and their responsibilities, such as executive managers, policymakers, AI developers, and consumers, many of whose responsibilities have been complicated by AI.
  • Found that current AI literacy research covers responsibilities like identifying AI opportunities but has significant knowledge gaps in areas such as "identifying AI benefits," "estimating AI costs," "strategic planning," and "monitoring and optimizing deployed AI systems."

• What advantages does it have compared to existing solutions?

  • Expanded the scope of AI literacy research to include regulators, organizational decision-makers, and those affected by AI, beyond just users and developers.
  • Provided a more detailed role classification, facilitating the development of tailored AI literacy programs for different responsibilities.

• What were the experimental or evaluation results?
  Key findings include:

  1. Current AI literacy frameworks do not adequately cover several critical responsibilities, such as evaluating AI's economic or social benefits and risks, as well as knowledge gaps related to strategic planning.
  2. There is also a lack of systematic guidance for monitoring and optimizing the impact of AI systems.

• Limitations and Future Directions

  • Limitations:
    • The service design approach leans towards a commercial perspective, potentially overlooking responsibilities of non-commercial roles (e.g., educators).
    • The literature analysis relies on existing news reports and research, which may not fully capture the complexity of real-world responsibilities.
  • Future Directions:
    • Develop professional development resources, such as AI literacy training for managers, policymakers, and other non-developer roles.
    • Collaborate with the Responsible AI community to explore knowledge gaps in areas like AI impact monitoring and strategic planning for risks and benefits.

Conclusion

Through service design and AI ecosystem role analysis, this study reveals significant, under-researched AI literacy needs and provides a framework to address these needs. The research encourages the expansion of AI literacy and cross-disciplinary collaboration, advocating for the development of educational resources tailored to non-user and non-developer roles to address the societal challenges posed by AI technology.