Designing for Responsible and Empowering GenAI Research, Policy, and Practice: Young Adults with Special Education and Neurodiversity Experiences Meaningfully Navigating Risks and Opportunities Toward Expansive Possibilities

Abstract

Generative AI (GenAI) has become what Engeström (2009) refers to as a “runaway object”, rapidly expanding across education, policy, and society in ways that are difficult to predict or control. Its growth creates both opportunities and challenges. Runaway objects demand collective responsibility and inclusive engagement in shaping how they impact society and how it is used. As a runaway object, designing responsible and empowering approaches to GenAI in education, must engage those who have been historically excluded from research and decision-making.This dissertation focuses on young adults with special education and neurodiversity experiences, a population often excluded from educational research and decision-making about new technologies (Rice & Dunn, 2023). These individuals bring valuable perspectives shaped by navigating systems that have not always met their learning needs, and by often developing deep relationships with assistive technologies that support their communication and learning (Dieker & Zaugg, 2024; Kleekamp, 2021). Using a design-based research (DBR) approach, this dissertation examines how young adults with special education experiences engage with GenAI. While DBR supports flexible, participant-responsive design, it can also reproduce power imbalances that limit full participation (Vakil, 2016). To address this, I adopted a collaborative co-design process, working with two to three participants from each group to make the research tools more accessible and inclusive. Data from co-design meetings, focus groups, and interviews were collected online through Zoom and surveys through Google Forms. Following the two cycles of coding model (Saldaña, 2016), I created codebooks through an inductive and deductive coding approach, then synthesized the first-cycle codes and identified broader themes. Data is analyzed through theoretical lenses based on Cultural Historical Activity Theory (CHAT; Engeström & Sannino 2010; 2020) explore how learning unfolds when young adults and adults work together to design responsible and powerful GenAI research, policy, and practice. CHAT allows me to examine co-design activity systems and how they are shaped by tools, rules, histories, contradictions, power dynamics, and the shared object that organizes the work. CHAT gives me the language to identify moments of tension, discussion, and expansive learning as participants reimagine their future of engagement with emerging technologies. The first article focuses on co-designing the research process, the second on co-designing GenAI policies, and the third on co-designing GenAI practices. Together, these studies illuminate the emotional and relational dimensions of GenAI engagement. This dissertation offers both practical insights and theoretical contributions for researchers, educators, and policymakers seeking to build more responsible and empowering GenAI supported learning environments.