Just and Ethical AI Educational Research and Development: A framework and design approach

• In Event: Generative AI in STEM Teaching and Learning: Informing Transformative Uses Through Ethical Research and Development

Abstract

Advancing personalized learning has been an engineering grand challenge for over a decade, inspiring the creation and research into data-driven artificial intelligence (AI) and advanced learning technologies. The emergence of generative AI (genAI), such as large language models like ChatGPT, has made the creation of new personalized learning technologies more accessible than ever before. However, AI technologies reflect both the intended and unintended biases of their designers and the wider society, making it imperative to adopt policies and practices that prioritize ethics, equity, and justice in research and development for education. Responsible AI research and implementation will require new frameworks and practices that allow for multi-disciplinary, multi-stakeholder discourse and interrogation into the impacts of AI in education. Guiding questions for inclusive and equitable AI in education can help multidisciplinary teams adopt continuous and reflective practices to focus on who benefits from AI, how people are prioritized, and why we believe that AI will improve education. The answers to these questions must be informed and discussed from multiple perspectives, from researchers to communities, educators, families, and students, and throughout the lifecycle of AIED projects and technologies. Furthermore, each AI design and usage decision must be informed with AI and data literacy that acknowledges the ways that AI and algorithms can harness, transform, and share data, and how these choices can impact people’s lives. Responsible, inclusive and equitable AIED research and development will need to become more transparent and explainable, and designed to counter oppression, while improving people’s AI literacy and autonomy.


This presentation will introduce the author’s ethical AIED framework grouped by the overarching principles of justice, respect, and beneficence, adapted to guide ethical AI research (Greene et al., 2024) and for education (Roschelle et al., 2024). For each principle, the framework provides a definition, a guiding question, and a list of related AI ethics principles and considerations for educational contexts. Additional principles were adapted from a review by Roschelle and colleagues (2024) of state-level guidelines for the use of AI in schools, including equity, inclusion, pedagogical appropriateness, and AI literacy. In addition, the framework highlights social, cultural, community, and societal dimensions of justice, rights, and educational roles.

This ethical AIED framework is meant as a tool for researchers and developers to help guide design priorities. The presentation will show how the framework can give rise to an ethical AI design reflection map which aims to provide explicit reminders of the relevant ethical considerations for designing and deploying an educational AI innovation. Finally, the presentation will show how the resulting map can be used both to guide the iterative design process and as a communication tool to structure a dialogue between stakeholders around claims made regarding educational outcomes and the inherent risks of such an approach.

Authors

• Tiffany Barnes, North Carolina State University

• Joshua Adam Danish, Indiana University