SIDLIT 2026

As generative AI becomes increasingly embedded in online and blended education, instructors face a difficult question: how can learning be supported and assessed when students can quickly generate polished answers, explanations, lesson materials, and multimedia products? Much of the current conversation focuses on academic integrity, prompt writing, tool adoption, or AI-use policies. While these issues are important, they do not fully address a deeper instructional question: what does meaningful learning look like when AI-generated output becomes part of the learning process? This work-in-progress session begins with a dialectical autoethnographic inquiry into a 24-turn interaction between the presenter and a generative AI system. Although the interaction began with a personally meaningful problem related to IRA planning, the focus of the analysis is not financial decision-making. Rather, the episode is used as a situated case for examining how learning unfolds within a human–AI–artifact system. Preliminary analysis suggests that AI-generated outputs should not be treated as final answers, neutral tools, or authoritative explanations. Instead, they function as epistemically unstable learning materials that require human verification, revision, justification, and synthesis. Building from this analysis, the session translates the emerging theoretical insight into practical strategies for online and blended teaching. The presenter will introduce assignment and assessment structures that foreground process evidence rather than only final products, including prompt archives, revision logs, delta reports, AI feedback loops, AI defense activities, peer process audits, and reflective synthesis prompts. These strategies are designed to help instructors evaluate how students define problems, examine AI-generated output, verify information, revise their thinking, justify decisions, and produce a defensible final synthesis. Participants will leave with a conceptual vocabulary for understanding generative AI-mediated learning and a set of adaptable strategies for designing assignments that make student judgment visible. The session is intended for educators, instructional technologists, online program leaders, and educational technology researchers interested in moving beyond AI-use compliance toward more rigorous, reflective, and assessable forms of AI-supported learning.