- What is the Statement of Shared Practice?
The Statement of Shared Practice is a voluntary, 12-month commitment among institutions that steward archival and special collections. It establishes shared principles for evaluating new requests from companies seeking access to these collections for AI model training. The issue is a current one: AI companies are approaching research libraries now, and most institutions are evaluating these requests individually, without shared language or visibility into what others are being offered. The Statement changes that.
- Does this prevent institutions from engaging with AI?
No. The Statement favors AI approaches that preserve the traceable connection between the system's outputs and the source materials. Institutions retain full authority to approve any request they choose.
The Statement establishes a floor for responsible practice, not a ceiling.
- What collections does this apply to?
The Statement applies to archival, manuscript, rare, and special collections only. These materials are often held under donor agreements and custodial obligations with specific terms and expectations. Many were entrusted by donors, communities, and predecessor organizations for preservation and scholarly access. The Statement does not apply to general library collections or widely held published works.
- Does this conflict with open access principles?
The Statement addresses an archival question, not a library access question. Archives hold materials entrusted by donors, communities, and predecessor organizations under specific terms. When an institution evaluates whether to allow a new use of those entrusted materials, that is a disposition decision, not an access decision. Donors who entrusted their materials could not have contemplated AI model training. Evaluating a new use against the terms of that trust is a fiduciary responsibility. The Statement explicitly preserves every institution's ability to advocate for fair use and other legal doctrines.
- What about existing digitization partnerships?
The commitments apply only to new and prospective requests. They do not retroactively govern existing contractual relationships or legacy digitization partnerships.
- What is the relationship between the Statement and the UVA Archival AI Protocol?
The Protocol is the University of Virginia Library's institutional framework for evaluating AI training requests. The Statement translates the Protocol's principles into a collective commitment any institution can endorse. Institutions sign the Statement, not the Protocol. The Protocol and its adoption guide and implementation toolkit are available as open implementation resources under CC BY 4.0 for any institution to adapt.
- What does endorsement obligate my institution to do?
Signing on to Shared Practice is a professional commitment, not a legal contract. It signals that your institution shares these principles and intends to apply them when evaluating new AI training requests for your collections. There are no penalties for imperfect implementation. Institutions may withdraw at any time by written notice.
- Is this an agreement to restrict access or coordinate against AI companies?
No. Each institution makes its own decisions independently. Nothing in the Statement constitutes an agreement to refuse to deal with any party or to set prices or terms.
The shared ledger — an anonymized record of types of requests and terms received by participating institutions, offered to signatories of the Statement — explicitly excludes pricing, financial terms, and compensation structures.
- My institution is being approached by an AI company right now. How does this help?
Endorsement of the Statement connects your institution to a community of peers navigating the same challenges.
- The Statement gives shared definitions for classifying the request and a framework for evaluating provenance and control.
- Signatories will have access to a shared ledger of confidential, anonymized records of the types of requests participating institutions are receiving and the general structure of terms proposed, thus providing visibility into how peers are approaching similar requests.
- The UVA Archival AI Protocol implementation guide and adoption toolkit offers practical tools including a decision framework and model contract language, available under a CC BY 4.0 license.
- Who has endorsed?
The founding cohort of signatories includes:
- University of Virginia Library (coordinating institution)
- Duke University Libraries
- Florida State University Libraries
- Northwestern University Libraries
- Oklahoma State University Library
- Rice University, Fondren Library
- Tulane University Libraries
- University of Florida, George A. Smathers Libraries
- University of Rochester, River Campus Libraries
- University of Wisconsin-Madison Libraries
- Washington University in St. Louis Libraries
- Wayne State University Libraries
The Statement was developed following an ARL peer-to-peer session attended by participants from more than 42 institutions, and revised based on feedback from founding cohort members, institutional counsel, and colleagues who engaged critically with the original draft.
- Can my institution join after the founding cohort?
Yes. Endorsement is open to any institution that stewards unique cultural collections throughout the 12-month period.
- What happens after 12 months?
The Statement expires on April 3, 2027, unless participants affirmatively renew it. The community will reconvene by February 2027 to assess what was learned and whether to continue, revise, or conclude.
- How do I sign the Statement of Shared Practice?
Endorsement requires the signature of the institution's senior leader (dean, director, university librarian, state archivist, or equivalent) and takes about 60 seconds. Use this form to register your endorsement.
