When AI Comes for the Library

The proposal

Overview

Artificial intelligence is rapidly becoming one of the primary interfaces through which people search, synthesize, and interact with knowledge. It draws on intellectual infrastructure that universities, libraries, publishers, and researchers have spent generations building.

This paper confronts a simple question: What if trusted knowledge institutions built their own living knowledge architectures that AI could consult, rather than simply allowing AI to consume their intellectual infrastructure?

An outside AI could bring a question to an institution's digital gateway. Behind that boundary, institutional knowledge, experts, agents, methods, and validation systems could produce a traceable consultation. The AI could receive an answer, supporting evidence, and stated limitations without receiving unrestricted access to the architecture that produced them.

Many of the technical components already exist. At the ISKO 2026 conference in São Paulo, I saw them developing separately: machine-readable knowledge organization systems, ontologies and knowledge graphs, research workflows, and approaches to provenance, transparency, and auditability. The harder challenge may be institutional: whether fragmented universities can align their systems, governance, methods, and people well enough to operate as coherent knowledge architectures.

This is intentionally an optimistic proposal. It imagines universities and libraries remaining active, sovereign participants in an AI-mediated knowledge economy by making human stewardship, institutional trust, and distinctive ways of knowing the center of value.