Responsible AI in Libraries and Archives
An IMLS-funded project aiming to support ethical decision-making for AI projects in libraries and archives.
Responsible AI in Libraries and Archives (2022-2025) produces tools and strategies that support responsible use of AI in the field. AI projects in libraries and archives can support increased impact and new uses of resources, and Responsible AI provides new data, new resources, and new strategies that will prepare our profession for the methodical consideration of potential harms of AI projects. As we begin to ethically and responsibly operationalize AI in libraries and archives, we hope that libraries and archives can provide models that resonate beyond the profession, embodying Library of Congress Director of Digital Strategy Kate Zwaard’s idea that “through the slow and careful adoption of tech, the library can be a leader.”
Project Updates
March 2024: Now published! Journal of eScience Librarianship Volume 13, Issue 1, Special Issue: Responsible AI in Libraries and Archives
The Journal of eScience Librarianship (JeSLIB) and the Responsible AI in Libraries and Archives project are proud to publish a Special Issue on Responsible AI in Libraries and Archives.
Librarians and archivists are often early adopters and experimenters with new technologies. Our field is also interested in critically engaging with technology, and we are well-positioned to be leaders in the slow and careful consideration of new technologies. Therefore, as librarians and archivists begin using artificial intelligence (AI) to enhance library services, we also aim to interrogate the ethical issues that arise. The IMLS-funded Responsible AI in Libraries and Archives project aims to create resources that will help practitioners make ethical decisions when implementing AI in their work. The case studies in this special issue are one such resource.
Guest editors: Sara Mannheimer, Doralyn Rossmann, Jason Clark, Yasmeen Shorish, Natalie Bond, Hannah Scates Kettler, Bonnie Sheehey, and Scott W. H. Young
The full issue is available at https://publishing.escholarship.umassmed.edu/jeslib/issue/59/info.
September 2023: JeSLIB special issue announcement, visioning workshops
We are excited to announce that we will be partnering with the Journal of eScience Librarianship to publish a Spring 2024 special issue featuring our responsible AI case studies. The issue will also include an in-depth literature review of AI projects in libraries and archives, focusing on how ethical issues are addressed in the literature.
During Fall 2023 and Spring 2024, we will also host a series of Responsible AI visioning workshops — some with librarians & archivists, and others with library users. These workshops will inform our ethical decision-making tool by helping us understand the ethical issues facing practitioners and users.
August 2023: Welcome, Natalie Bond!
Natalie Bond, the Government Information Librarian & Head of Information and User Services at University of Montana, has joined the grant as a co-Investigator. Natalie will provide a reference and instruction perspective.
March 2023: Case studies in progress
The Responsible AI team is working with a cohort of eight author groups to create case studies exploring AI-related ethical challenges in libraries and archives. These case studies will aid in the development of an ethical decision-making tool.
The case studies will be formatted according to the template below.
Responsible AI Case Study Template
Please use simple, accessible language to describe your case study, organizing your narrative according to the following six prompts. Case studies should be no longer than 10 pages (text of no more than 5,000 words).
1. Summary
Provide a 4-5 sentence summary of the project.
2. Project details
Describe what you did;
- who was involved
- what their role(s) were;
- what services were drawn upon;
- what collections were involved and why were they selected;
- what infrastructure and technologies were selected and why;
- what challenges were encountered in the course of the implementation?
3. Background
Describe why the decision was made to implement (or why there was an attempt to implement) the AI project/tools in your library and/or archive.
- What are the benefits of the project or tool?
- What gaps or problems are you trying to address?
- What other tools or projects did you look to for guidance or inspiration?
4. Ethical considerations
Describe any ethical issues that arose as you implemented the project, potentially including:
- How you considered and addressed potential harms associated with the implementation.
- Privacy considerations
- Do users consent to data sharing and use?
- Did you engage with any stakeholders?
- Did you refer to existing documentation, policy, or best practices?
- Did you refer to the ALA Code of Ethics or other codes?
- Did you conduct risk-benefit analyses?
- Did you hear concerns from your user community or your library?
- Include any considerations that are considered unresolved.
5. Who is affected by this project?
Describe the people, services, and programs that are involved in, support the use of, or are affected by the AI implementation (e.g. digital scholarship services, metadata services, subject area liaisons, library users, etc.)
6. Lessons learned and future work
Describe any lessons learned regarding the ethical and responsible implementation and use of AI. What recommendations would you give to others pursuing similar work? Include here any unanswered questions that merit further investigation, and/or any plans to build on the work described in the case study.
7. Documentation
Include any formal documentation (personas, use cases, code samples, functional requirements, privacy policies) or workflows that support the implementation.
Responsible AI Project Overview
Project Timeline
|
Phase
|
Dates
|
Activities
|
Deliverables & Dissemination
|
|---|---|---|---|
|
Phase 1. Environmental
Scan |
Aug 2022 - |
|
|
|
Phase 2.Case studies&beneficiary |
Mar 2023 - Sep
2023 |
|
|
|
Phase 3a. Target group |
Oct
2023 - Mar 2024 |
|
|
|
Phase 3b. Ethically-relevant
Harms Analysis Tool Assessment and Validation |
Apr 2024 - Oct
2024 |
|
|
|
Phase 4. Final Ethically- |
Nov 2024 - July
2025 |
|
|
Research Outputs
All research products are available on the Open Science Framework: https://doi.org/10.17605/OSF.IO/RE2X7
Project Team
Project Director
- Sara Mannheimer is Data Librarian at Montana State University (MSU). She has experience teaching and and conducting computational data analysis using the R programming language. Her research focuses on quantitative and qualitative social science research methods and applied ethics. She is affiliated faculty in the MSU Center for Science, Technology, Ethics, and Society and serves as project manager for the MSU Dataset Search.
Co-Project Director
- Jason Clark is head of the Research Optimization, Analytics, and Data Services department at MSU Library. He is a technologist and software application developer (PHP and Python) whose work focuses on innovations in library software, including AI applications such as speech-to-search and anticipatory design search. Clark was PD of the IMLS-funded Algorithmic Awareness project, which created teaching resources to support understanding of the practical and ethical implications of algorithms, including algorithms for AI.
Co-Investigators
- Doralyn Rossmann is Dean of the Library at MSU. Her background is in information technology in higher education and libraries. Rossmann has managed IT systems and software development at several universities and has served on MSU’s Senior Information Technology Leadership Team. She has expertise in writing privacy policies and management of user data in libraries. Her research focuses on social network data ethics and augmented and virtual realities.
- Bonnie Sheehey is an ethicist in the MSU Department of History and Philosophy. Sheehey focuses on the social, ethical, and political implications of technology. She has published on issues of racial bias and injustice in current technologies deployed in a variety of criminal justice practices in the U.S. context. Sheehey is affiliated faculty in the MSU Center for Science, Technology, Ethics, and Society.
- Scott Young is User Experience and Assessment Librarian at MSU. He is a library service designer and has experience facilitating multi-day participatory design workshops similar to that being proposed above. He served as PD of the IMLS-funded National Forum on Web Privacy and Web Analytics (co-PDs Mannheimer and Clark), which featured a 2.5-day forum thatfunctioned as a participatory design workshop in support of privacy-oriented library services. His research focuses on applying library values through the lens of practical ethics. Young is affiliated faculty in the MSU Center for Science, Technology, Ethics, and Society.
- Natalie Bond is Government Information Librarian & Head of Information and User Services at University of Montana. She is interested in human rights and social justice as they relate to archives and library services.
- Yasmeen Shorish is Director of Scholarly Communications Strategies at James Madison University. Her work focuses on data ethics and privacy. With expertise in data librarianship and scholarly communications, Yasmeen has a unique perspective on how data products and computing intersect with the scholarly record. She has presented and led workshops on the threats posed by digital surveillance on privacy and society. She has also investigated issues related to data discoverability in her leadership on the IMLS-funded OA in the Open and ethical means of computational analysis of library collections in her involvement with the Mellon-funded Collections as Data: Part to Whole project.
- Hannah Scates Kettler is Associate University Librarian for Academic Services at Iowa State University. Her work focuses on creating and highlighting intervention points in library work that invite critical assessment of approach and outcomes. Through her work in the DLF Cultural Assessment Working Group, she has co-developed intervention points for selection and appraisal of digital collection creation and preservation and instills those concepts at the local level. She has worked on the development of community driven 3D data preservation standards, shepherding digital collections as data models which focus on ethical use and reuse through her involvement in the IMLS-funded Collections as Data: Always Already Computationalproject and leadership in the Mellon-funded Collections as Data: Part to Whole projects.
Advisory Board
- Dorothy Berry, Smithsonian Institution
- Stephanie Russo Carroll, University of Arizona
- Bohyun Kim, University of Michigan
- María Matienzo, Tome
- Thomas Padilla, Internet Archive
Project Sponsors
This project is made possible in part by the Institute of Museum and Library Services, through grant # LG-252307-OLS-22. The IMLS is the primary source of federal support for the nation’s libraries and museums. They advance, support, and empower America’s museums, libraries, and related organizations through grantmaking, research, and policy development. Their vision is a nation where museums and libraries work together to transform the lives of individuals and communities. To learn more, visit www.imls.gov.