The open education movement builds on the human desire to collaborate, to share, to guide other humans. Educators have harnessed those impulses to grapple with AI in education through open educational practices and OER about AI. And AI has begun to help us with OER discovery, curation, review, fact-checking, and dissemination. It helps us customize and offer more modes of learning. Dialogue with AI can stimulate fact-checking and revision. But at the same time, when we let AI do too much, we risk alienating students and abandoning the social strengths of open. This talk will explore how we can develop shared norms for limits and transparency around AI use in OER. Guidelines can encourage us to convey core concepts primarily through human words and ideas. Templates for disclosure and process history can help us build trust, hold ourselves accountable, and model wiser choices around AI use. We will examine gray areas within case studies and share perspectives on where to draw the line and how to communicate to students our commitment to education as human guidance.
Pockets of open ed spent the first wave of generative AI learning to prompt carefully and evaluate outputs, while also asking how generative OER might extend the older work of adopting and adapting around the 5Rs. Agentic AI, or GenAI 2.0, shifts the focus from generating content in a chatbot interface to delegating work across a broader OER ecosystem. That ecosystem may include textbook material, companion websites, GitHub repositories, source collections, APIs, and local materials an OER creator wants to connect through agent platforms like Claude Code. We are already seeing experiments with new OER bundles. Some pair traditional textbook content with websites, code, data, update scripts, and maintenance practices the creator may only partly understand. Will these workflows quietly erode attribution and flood the commons with yet more AI slop? How can they be designed around stewardship and the open ethos that brought us here? This talk explores the new OER possibilities opened by agentic AI and offers a practical framework for navigating them in ways that refocus attention on human judgment and creativity. Paradoxically, the agentic moment may make original, human-authored OER more valuable, because agents are most useful when they extend and sustain work of an individual or community that remains deeply invested in it.