- American University’s two-day workshop runs on 26 and 27 October.
- The organisers encouraged each institution to send a senior leader and a collections or digital stewardship practitioner.
- A small trial can specify who reviews the work, what counts as an error and who can stop the process.
Inside this story 2 perspectives
Take two people to a meeting about changing the way a library works. Give one of the places to someone who can commit the organisation’s time or money. Give the other to someone who will have to handle the material on a Tuesday morning after the decision takes effect. A workshop at American University in Washington has built that pairing into its invitation.
The programme, scheduled for 26 and 27 October, brings together American University, Hofstra University and JSTOR Digital Stewardship. Organisers encouraged institutions to send a library leader and a colleague from special collections, archives or digital stewardship. They have designed the sessions around organisational change, including the introduction of AI tools into work that depends on human judgements about context.
Put a real object on the table
An archival photograph makes the discussion less abstract. In an October update, JSTOR described a collection of photographs from Hempstead, New York, held by Hofstra. Its description tool helped process the collection and Hofstra staff reviewed the results. One image shows a hotel and street corner, with smaller images arranged around it. A computer-generated description that identifies a building may still leave an archivist to determine a name, a location or the relationship between the pictures.
A team considering a similar change could start with one photograph and write down exactly what would have to happen before its description could be made public. Someone must check a proposed identification against the records. Someone must decide what to do when a building cannot be confidently named. A manager estimating the time saved needs to see both the straightforward item and the one that sends a colleague back to a box or a catalogue.
Agree who can pause the work
A small trial becomes easier to interpret if the people involved decide beforehand what would count as a problem. An error in punctuation and a confident misidentification of a person have different consequences. A reviewer needs a place to record uncertainty and authority to hold the affected record back. The person leading the project needs to know whether resolving one difficult record took two minutes or an afternoon.
The workshop agenda includes building a charter for change. Even a short version can say who owns a decision, who carries out the checks and when the team will meet again. The pairing gives each person something specific to bring back from that meeting. The practitioner can identify which part of the new work needs attention. The manager can decide whether the institution has allowed enough time and authority to do it properly.



