1 of 30

Navigating AI Integration: �A Survey of Big Ten Academic Alliance Libraries

Win Shih

USC Libraries

eCAUG Meeting, 10/31/2025

2 of 30

Background

Accelerated AI product launches
Establishing a proactive library process
Strategic opportunities
AI Task Force

Clarivate. (2025, October 30). Pulse of the library 2025.

First, I will provide a background of our project.
AI is here to stay. According to Clarivate’s Pulse of Library 2025 report, released 10/30/2025, 67% of survey participants indicated that their library are exploring or implementing AI technologies, comparing to 63% in the 2024.

Background (When the landscape keeps shifting, relying on old ways for operations and decision-making quickly become outdated and ineffective.)

Accelerated AI product launches (The catalyst in the AI vendor landscape): During the fall of 2024, several library vendors rapidly launched AI-powered products, often with minimal advance notice and user needs assessment (Primo RA was introduced on 9/9/2024).
Establishing a more proactive, structured, coordinated, and cross-functional process to work with our leadership in

Systematically assessing and veting new AI tools
Formulating official recommendation s
Developing comprehensive staff and patron training programs
Engaging with vendors on product details

Strategic opportunities: Recognize a chance to propose a wholistic framework for the adoption and implementation of AI technologies

AI Task Force formation and charge – in response

Formation: In response and with support from the AD of Library Technology, the USC Libraries AI Task Force was formed in November 2024
Leadership: Co-chaired by myself and our Head of Instruction
Composition: Comprises 8 members from key functional areas (public service, instruction, health sciences library, technical services, and IT)
Primary charge: To develop formal principles for AI adoption and to foster a community of practices around this emerging technology.

3 of 30

Project Overview

Project objective: Understand AI adoption across 19 BTAA libraries
Methodology: a Qualtrics survey distributed to IT leads at BTAA libraries
Period: 1/22 – 3/16/2025
Response: 12 responses (63%)

4 of 30

Big Ten Academic Alliance

A collaboration between Big Ten university libraries
Founded in 1961
Advancing the academic excellence of member institutions through national leadership, collaboration, and co-investment
19 academic libraries

5 of 30

6 of 30

Survey Questions

AI Strategies and Governance
AI Use Cases, Projects, and Applications
AI Literacy and Professional Development
Future Outlook

7 of 30

Scope

8 of 30

I. Leading - Strategies & Governance�

9 of 30

AI Policies

73%: No formal AI policies
9%: Policies in development
Quotes:

“We’re developing statements on AI use with collections and aligning with university policies”

“We have an AI Advisory Group that is developing ‘statements’ about the use of AI with our collections and reinforcing current University AI and Data policies”

10 of 30

Strategic Alignments

AI in library documents:

50%: mentioned in other initiatives or documents (e.g., annual reports)
40%: featured in strategic plans or mission statements

Quotes:

AI is a pillar in our strategic plan for discovery and academic innovation
Planning a Year of AI for the upcoming academic year
Using AI for ADA compliance with Title II regulation
Ongoing ML project to generate metadata
Build collections and infrastructure to support specialized research

11 of 30

Institutional Preparedness

“Libraries have begun exploring [AI] vendor tools and drafting guidance, but too often without comprehensive policies, clear principles, or frameworks that address equity, sustainability, and transparency.”

McBride, M. (2025, October 30). Do academic libraries have a strategy for AI? The Scholarly Kitchen.

12 of 30

Leadership & Decision Making

AI leadership is distributed:

Library IT
Digital Initiatives
Technology and Digital Programs divisions
Joint leadership between Digital Scholarship and Research/Learning units

Decision-makers varied by project size

Senior leadership
Dean, UL, AUL
Input from executive teams
Distributed

13 of 30

Budget and Resources

50%: Partial funding within existing technology budget
50%: No dedicated budget
Some rely on internal/external grants

Question: Do you have a dedicated budget for AI-related projects?

How the library accommodate new AI-related costs

Half of the respondents said they do not have a dedicated funding for AI-related costs
Another half indicated they do not have defined budget line for AI. Rather they reallocated existing budget to accommodate AI spending
In open-ended comments, respondents identified grants as another possible funding source
The implementation of AI initiatives requires dedicated investment across critical areas: financial resources, specialized personnel, and technology infrastructure. Our BTAA survey shows that 50% of libraries lack dedicated budget for AI projects. In addition, 20% of libraries secured funding through internal or external grants. To ensure sustainable AI adoption, we recommend establishing recurring budget allocation, integrating AI costs into standard IT expenditure planning, and pursuing strategic grant opportunities.

14 of 30

Staffing

AI Positions:

40%: No dedicated AI positions
30%: Created AI roles
20%: Planning
Avoid creating a dedicated position

Positions:

Machine Learning Engineer
AI Librarian
Lead AI Developer

15 of 30

Collaborations

100%: Collaborate with on-campus partners

Central IT, high performance computing, digital repository
Teaching centers
Academic units, colleagues

40%: Work with external partners (33%)

Consulting firms, vendors
Mellon Grant collaborations

16 of 30

II. Operating & Innovating��

17 of 30

AI Applications in Use

Research and Discovery

67%: AI-powered search/discovery tools

Primo Research Assistant, Scopus AI, Semantic Scholar, WoS AI

60%: Research support and literature review tools

Connected Papers, Research Rabbit, Scite.ai, Dimensions.ai

Metadata Management

50%: Cataloging enhancement tools

Alma Metadata Assistant, Annif

User services

Chatbots

18 of 30

AI Applications in Use

General Purpose Tools

Microsoft Copilot, ChatGPT, Google Gemini, Claude, and Perplexity

Home grown

Developed a semantic search tool for digital collections
Digital publishing workflows
AI literacy initiatives

19 of 30

Evaluation Practices

Methods:

75%: Pilot testing and evaluation
67%: Cost-benefit analysis
50%: Privacy and security risk assessment
42%: Performance benchmarking and comparison
42%: Vendor presentation and demonstration

Rubrics:

40%: in development
20%: in use

Central IT performs accessibility and security assessments

20 of 30

III. Learning & Teaching��

21 of 30

AI Literacy

Patron Programs

85%: AI-related research guides and online resources
58%: Workshops on AI research tools
42%: Ethical AI use

Examples:

Hosting AI Meetup on campus
AI literacy brownbag series
Information literacy workshops in the age of AI

Question: What AI literacy programs do you offer to patrons?
Libraries are uniquely positioned to provide accessible educational program on AI through research guides, workshops, and tutorials. Beyond technical skills, such programs should foster critical discussions on AI ethics and responsible use, aligned with the ACRL AI Competencies for Library Workers. Libraries can also serve as hubs for open dialogue on AI’s societal challenges and opportunities. The BTAA survey indicates:
Some of the current AI-related offerings at USC Libraries are:

Developed comprehensive AI research guides
Conducted AI workshops on:

Generative AI in Research
AI Literature Mapping Tools

Created instructional tutorials
Hosted three well-attended AI meetups in Spring 2025 (averaging 35 participants per session) featuring discussions on:

Practical AI tool experiences
Ethical considerations
Educational and personal impacts

22 of 30

Staff Development

Examples:

Planning “Year of AI” events
Professional development funds for AI training
Town halls discussion on various AI implications and offering on-demand courses from LinkedIn Learning
Identifying AI leads in specific areas for departmental training
Monthly brownbag reading group
Informal workshops and presentations
Ongoing AI interest groups

Question: What AI-related professional development opportunities do you offer?
As the nature of work continues to evolve, so do the skills required to succeed in an increasingly automated, technology-driven world. upskilling or reskilling existing faculty or staff
Staff development investment: Libraries are actively expanding AI competencies through diverse staff development initiatives, as demonstrated by the BTAA survey:

Planning “Year of AI” events
Professional development funds for AI training
Town halls discussion on various AI implications and offering on-demand courses from LinkedIn Learning
Identifying AI leads in specific areas for departmental training
Monthly brownbag reading group
Informal workshops and presentations
Ongoing AI interest groups

To maximize the impact, libraries should:

Develop structured upskilling and reskilling programs on AI technologies
Support experimental integration of AI tools in workflows
Foster collaborative learning environment

23 of 30

IV. Outlook��

24 of 30

Major challenges

Funding
Staff capacity and skills
Ethical concerns
University policies and compliance
Evaluation of vendor products

25 of 30

Clarivate Pulse of the Library 2025

Budgets
Privacy and Security
Lack of expertise
Research and academic integrity

Clarivate. (2025, October 30). Pulse of the library 2025.

26 of 30

Future Direction

AI literacy
AI-enhanced research tools
Enhanced discovery of library content
Metadata and cataloging (backlog)
Accessibility
Public services

27 of 30

NISO Survey

Top three common hopes:

Improved visibility of content
Enhance the value of web scale discovery
More accurate content recommendations

Varnum, K., & Kessler, R., Zhu, J., Hazen, T., Patham, B., & Holloway, J. (2025 August). Generative artificial intelligence and web-scale discovery. NISO Open Discovery Initiative Standing Committee.

28 of 30

Innovative Projects

Northwestern’s semantic search for digital collections
Ohio State’s PDF Accessibility Initiative - AI for accessibility improvements
UMichigan’s copyright scenarios generated with AI
Annif for Subject Assignment for ETD
Automating metadata generation and cataloging processes
Mellon Grant-funded collaborations (e.g., NYPL, UT Austin)
USC Interactive Interviews – AI-driven service
AI-powered transcription and description tools
Developing AI literacy programs
Crating AI task forces
Chat evaluation tools

29 of 30

eCAUG / ELUNA Opportunities

Sharing best practices and use cases
Inter- and intra- organizational cooperation and collaboration
Training
Showcasing
Facilitating collaboration and networking (e.g., AI-SIG)

30 of 30

Questions?�

Dr. Win Shih

winyuans@usc.edu

University of Southern California

https://tinyurl.com/ecaug2025ws

Invitation to share your stories

AI is an disruptive technology and it is change constantly. When the landscape keeps shifting, decisions relying on our old assumptions and practices quickly becomes dated and ineffective. What worked in the past may no longer fit.
What’s the cost of waiting? in volatile environments, the pursuit of perfect clarity often conceals a hidden cost: the cost of inaction
Questions: What decision today will still make sense a year from now? What if this isn’t the storm, what if it is the climate? What is the cost of waiting?
I hope the insights we gathered provide some guidance in your planning of AI projects, ensuring alignment with both library and university mission and objectives.
If the library communities were to wait for certainty, it would not only be a false goal, but would lead to them losing the opportunity to determine how best they can advocate for and position further development of tools, methodology and guidelines to enable a thriving future in the research and knowledge ecosystem.