AI-assisted discovery for archival digital collections
I’m doing some informal research on what archivists and curators think about new agentic search tools like Northwestern’s AI mode for their digital collections platform. My organization is a research leader in the US, and interested in implementing AI to solve a number of discovery and access issues, but we also have culturally sensitive archival materials.
I find it scary that nonprofits are hot to trot for this technology and aren’t giving the environmental impacts a second thought, and of course there are the obvious issues of ethics, bias, colonialism / data sovereignty… the list goes on. We use a lot of open source software, and are part of a research consortium, so I’m also concerned about the responsibility to get group consensus, but I’m scared to air these concerns because our Chief is very eager to jump on the bandwagon and I don’t want to lose my job or get stuck on a “special project” while other colleagues get promoted.
So, I’m asking here to get the temperature. What is everyone feeling? If you were forced to let AI use your collection data, what would your primary asks be to ensure safety and integrity and trust in your archives?