IGeLU 2026

Conference

AI discovery tools are often presented as more intuitive and user-friendly than traditional search. But before making them available to patrons, libraries need to ask difficult questions: Do they return authoritative sources? Are results relevant? How do they handle biased or harmful language?

This presentation shares SLSP’s comparative testing of the Primo Research Assistant, Ask Anything and Basic Search in Primo. The evaluation assesses results according to five criteria: bias, authority, currency, relevance and availability.

The presentation will share the methodology, key findings and limitations of the test, including reflections on biased query language, source diversity, open access visibility and the need for recurring evaluation as discovery tools evolve.

User Experience / Discovery

Primo