Increasingly, scholarly research does not lead to an article or a journal issue. It leads to an answer. That shift creates a practical problem for libraries, publishers, researchers, and infrastructure providers. When an AI tool reads across the scholarly literature and produces a response, the original work may be hard to identify, authorized access may be hard to enforce, and usage may be hard to count. The issue is not whether AI belongs in scholarship, but whether AI-mediated use will remain visible, attributable, governed, and measurable. Charleston is one of the few venues where this problem can be discussed by the people who need to make it work: librarians, publishers, platform providers, and analysts who sit between policy and implementation.

A library can have a license, a publisher can have a policy, and an AI tool can still produce an answer in which the source is invisible, authorization is unclear, and usage disappears from the evidence base libraries use to make decisions. That is the gap this session addresses.

The panel, at the Charleston Conference 2026, is timely because the technical and policy rails are being laid now. Publishers, infrastructure organizations, and technical standards conversations are converging on a shared question: what should accountable AI-mediated access to scholarly content look like in practice? Charleston’s audience should not inherit those choices after they are made. Libraries should help shape them.

Join the session on Thursday, November 5 at 4:00-4:40pm.