Writing with AI: Toward Transparent, Human-Centered Learning Systems in the Age of Generative AI

The rapid emergence of Generative Artificial Intelligence (GenAI) has introduced new possibilities and tensions for education, challenging established assumptions about authorship, feedback, literacy, and assessment. This talk presents a synthesis of a multi-year research program (2023–2026) examining the integration of AI in educational contexts through both empirical investigation and iterative system design.

Drawing on convergent mixed-methods studies across higher education and K–12 settings, this work traces the evolution of AI-supported learning environments, from early implementations of AI-generated feedback and AI literacy development to more recent design-based approaches that integrate pedagogical frameworks, analytics, and human–AI collaboration within digital platforms. Across these studies, findings indicate that while GenAI can provide scalable, rubric-aligned, and actionable feedback, its educational value depends on learners’ capacity to interpret, evaluate, and apply feedback within structured instructional contexts. At the same time, AI-supported environments can foster advanced digital literacies, multimodal meaning-making, and more inclusive forms of participation when grounded in pedagogical designs such as multiliteracies and translanguaging.

Building on these insights, the talk examines a human-centered design perspective operationalized through the ongoing development of CyberScholar, an AI-supported learning and knowledge environment that integrates rubric-based feedback, knowledge-base–grounded AI support, peer and AI review workflows, and learning analytics to scaffold student writing and evaluation processes. Designed to align AI outputs with pedagogical goals, the system supports structured, iterative engagement with feedback and positions learners as active participants in knowledge construction. As part of this work, emerging approaches to documenting and reflecting on AI use in the writing process are being explored, with the aim of making human–AI collaboration more visible, traceable, and pedagogically meaningful.

The talk advances the concept of Human-Centered AI Transformations (HCAIT) as both a research agenda and a design framework, emphasizing transparency, co-agency, and pedagogically grounded AI integration. It invites discussion on how such approaches can reshape understandings of authorship, assessment, and AI literacy in education.

Photo by Theo Crazzolara on Unsplash