Keeping a Human in the Loop: Optimising Donation Workflows with Multi-Agent Workflows

Activity: Talk or presentation typesPresentation

Description

This paper presents the results of an experiment conducted at the University of St Andrews, where artificial intelligence was used to optimise the processing of a backlog of donated books. The primary goal was to develop a multi-agent workflow that integrates AI to streamline decision-making, enhance operational efficiency, and handle large volumes of data, all while keeping human oversight at critical decision points to ensure control and ethical standards. This paper argues that a human-in-the-loop model offers libraries a way to continue to embed an ethical approach to our work, while also harnessing the improvements AI can offer our workflows.
Period20 Mar 2025
Held atResearch Libraries UK, United Kingdom
Degree of RecognitionNational