AMPlifying AV: Next Steps for the Audiovisual Metadata Platform

Since 2018, the Indiana University (IU) Libraries have been working with partners at AVP, New York Public Library (NYPL), and the University of Texas at Austin, with support from The Andrew W. Mellon Foundation, to build the Audiovisual Metadata Platform (AMP). AMP is an open-source software platform that allows archivists and librarians to define and execute custom workflows that combine artificial intelligence and machine learning services with human expertise in order to more efficiently and effectively create metadata for audio and moving image resources in support of discovery, identification, navigation, and rights determination. This session will share lessons learned from the last phase of development work and provide an update on the current phase of AMP work, including packaging the platform for broader implementation, experimental use of the system by librarians and archivists at IU and NYPL across a broad range of collections, and the development of an evaluation tool that allows collection managers to compare multiple tools and/or machine learning models to select the best one for a given task in a particular collection.