UKSG webinar: The Open Access – AI Conundrum: Does Free to Read Mean Free to Train? Feb 05, 2026, noon (GMT) | UKSG

The Open Access movement, originally designed to democratise human access to research, now faces unintended consequences as AI systems extensively use freely available academic content for training large language models. While Creative Commons licenses permit such use, researchers primarily chose open access to maximise human readership, not to provide free training data for commercial AI companies. This creates significant risks for academic ecosystems, including “citation laundering” where AI outputs obscure original sources, disrupting knowledge attribution and academic career incentives. At the same time, excluding AI systems from OA research carries its own risks: as high-quality scholarship retreats once again behind paywalls, AI outputs may become even more inaccurate, more vulnerable to data poisoning, and further prioritise misinformation over evidence.

This talk highlights the need for urgent policy discussions to navigate this dilemma – how to protect the research system’s foundational principles of transparency, attribution, and knowledge traceability, while also resisting the extraction of scholarly labour, the increasingly unmanageable cost of OA transformational agreements on libraries, and the enclosure of reliable knowledge in an AI-mediated world.