Tracking the AI Evolution in Research Libraries: Findings from ARL’s Third AI Quick Poll

by Leo Lo , The University of New Mexico , and Cynthia Hudson Vitale , The Johns Hopkins University | March 12, 2025

In just two years, the conversation around generative AI in research libraries has shifted from early interest to strategic action. As AI capabilities advance, libraries are navigating the implications for research, teaching, and information stewardship. To better understand this evolving landscape, members of the Association of Research Libraries (ARL) have been conducting a series of quick polls to capture member perspectives, challenges, and strategies related to generative AI adoption.

The third and latest poll in this ongoing series, conducted in January 2025, provides insights from 40 ARL library deans and directors about how their institutions are integrating and responding to AI technologies. This poll builds on the findings from previous polls, each reflecting a new stage in AI adoption:

• April 2023 (20 responses)—A snapshot of early experimentation, marked by cautious optimism and a recognition of AI’s potential
• December 2023 (74 responses)—A leap in engagement, as confidence grew, exploration expanded, and libraries began identifying use cases
• January 2025 (40 responses)—A more strategic and measured approach, as AI moves beyond experimentation into operational and policy considerations

This third poll offers a timely update on how ARL member libraries are navigating the challenges, opportunities, and institutional pressures of generative AI in an era of rapid technological advancement. The findings provide a snapshot of where we are now and where we might be headed next in the evolving relationship between AI and research libraries.

View on Generative AI Potential

Optimism about generative AI is evident among respondents. Over a quarter (28%) described their outlook as “very positive,” envisioning significant enhancements to library services in the next year. The majority (63%) expressed a “somewhat positive” view, acknowledging the potential of AI while being mindful of challenges. Only 10% maintained a neutral stance, reflecting a general trend toward growing confidence in AI’s role within libraries.

These results indicate a broader acceptance of AI as a tool capable of transforming workflows and improving user experiences, even as uncertainties remain about its broader implications.

Exploring or Implementing Generative AI

Engagement with AI technologies shows steady growth. Nearly one-third of respondents (28%) reported that their libraries are actively implementing AI solutions. The largest group (53%) is in the exploratory phase, investigating potential applications, while 19% indicated plans to consider AI in the near future.

These results reveal a clear trajectory of increasing activity, with most libraries either actively deploying or exploring generative AI tools as part of their strategy.

Ethical Use Policies

Responses reveal that libraries are at varying stages of developing ethical policies for generative AI. A small group (9%) reported having formal policies in place and actively enforced, showcasing an advanced approach to ethical governance. Meanwhile, 41% of respondents are in the process of developing policies, demonstrating widespread recognition of the need for structured guidelines. Another 25% rely on informal guidelines, which could serve as a foundation for more formalized frameworks in the future. However, 19% of libraries have no plans to develop policies, and 6% remain unsure of their institution’s stance.

This diversity highlights the ongoing journey toward responsible AI use, with many institutions prioritizing ethics as part of their broader AI strategy.

Financial Investments in AI

Financial investments reflect libraries’ commitment to advancing AI adoption. The majority of respondents (75%) reported funding professional development to upskill staff and enhance AI literacy. Subscriptions to AI tools and technologies were also a significant investment (63%), enabling libraries to access cutting-edge resources. Other investments include hiring trainers or consultants (25%) and creating dedicated AI-focused positions (22%). Notably, 9% of respondents reported no financial investments in AI, indicating that some institutions remain at the periphery of adoption efforts.

These trends underscore libraries’ focus on both human and technological capital to drive successful integration.

Personal Professional Development in AI

Respondents indicated a strong interest in professional development related to generative AI. A significant 63% have engaged in self-directed learning, such as exploring online tutorials or experimenting with AI tools. Additionally, 47% have attended workshops, webinars, or conferences, reflecting a collective push to gain structured insights. Formal certifications remain rare, with only 6% of respondents reporting completion of such programs. A small group (3%) plans to pursue professional development in the future, while 13% expressed no interest in AI-related training.

These findings suggest that flexible, accessible learning opportunities are critical for empowering library professionals to navigate the complexities of AI technologies.

Open-Ended Questions

AI Partnerships

Respondents emphasized that collaborations are a cornerstone of successful AI adoption strategies.

• Internal Collaborations: Most libraries work closely with teaching and learning centers, IT departments, research offices, and campus-wide task forces to align AI initiatives with institutional goals. These partnerships support shared projects like AI-enhanced teaching tools and ethical AI governance frameworks.
• External Collaborations: Libraries partner with vendors and interdisciplinary research groups to gain access to advanced tools and expertise. Examples include collaborations with AI developers and international groups like ai4Libraries to explore applications in libraries and archives.
• Grant-Supported Projects: Funding from organizations like the US Institute of Museum and Library Services (IMLS) supports partnerships with cultural heritage organizations and other institutions, enabling libraries to innovate while addressing resource constraints.

Employee Attitudes

Library employees’ attitudes toward generative AI vary widely:

• Curiosity and Enthusiasm: Many staff members are eager to explore AI’s potential, viewing it as an exciting tool to enhance workflows and services. Structured initiatives like AI clubs have fostered positive engagement.
• Cautious Optimism: Some employees remain skeptical, voicing concerns about ethical issues, job security, and privacy risks. However, they express a willingness to learn as they gain more exposure to AI tools.
• Divergent Reactions: Responses reveal significant polarization within teams, with some employees actively experimenting with AI while others remain hesitant.

Structured communication and training programs are key to addressing skepticism and building confidence among library employees.

Changing Views on AI

Respondents reported diverse experiences that have influenced their perceptions of generative AI over the past year:

• Positive Shifts: Many respondents expressed growing optimism, citing successful pilot projects like metadata generation and transcription improvements. Personal experimentation with tools like ChatGPT also enhanced their understanding of AI’s potential.
• Persistent Concerns: Ethical and environmental challenges remain top of mind. Concerns about misinformation, transparency, and resource disparities temper enthusiasm.
• Early Exploration: Some respondents emphasized that AI adoption is still in its early stages, requiring sustained experimentation and investment. Collaboration and professional development were frequently mentioned as catalysts for progress.

Trends Across the Three Polls

Shift from Cautious Optimism to Confident Exploration

The first poll in April 2023 revealed cautious optimism, with 55% of respondents “somewhat positive” and 10% “very positive” about generative AI’s potential. By December 2023, confidence had grown, with 67% “somewhat positive” and 18%“very positive.” The January 2025 poll continues this trajectory, as 28% of respondents are now “very positive” and 63% “somewhat positive,” reflecting steady growth in confidence as libraries see tangible benefits from AI applications.

Increased Engagement with AI

In April 2023, most libraries were in the exploratory phase, and only 10% reported active implementation. By December 2023, active implementation rose to 12%, and exploratory efforts expanded to 60%. The January 2025 poll highlights further progress, with 28% actively implementing AI and 53% still exploring. These results indicate a maturing landscape where libraries are steadily moving toward integration.

Collaboration as a Key Driver

Across all three polls, collaboration has consistently emerged as a critical factor in AI adoption. Early responses emphasized internal partnerships, while later polls highlighted the importance of external collaborations with vendors, interdisciplinary groups, and funding agencies. Libraries are increasingly leveraging these relationships to scale their AI initiatives and address challenges collectively.

Key Takeaways from the Third Poll

1. Optimism Grounded in Experience

The findings reflect a growing maturity in how libraries perceive and use generative AI. Optimism is no longer driven solely by the potential of AI but is increasingly supported by tangible outcomes from pilot projects and applied use cases. Metadata generation, transcription services, and discovery tools are proving their worth, reinforcing the value of continued experimentation and scaling.

2. Collaboration as a Critical Enabler

Internal and external partnerships are central to libraries’ AI strategies. Respondents consistently highlighted collaborations with campus IT departments, teaching and learning centers, and external vendors as essential to advancing their initiatives. These partnerships provide the technical expertise, resources, and alignment necessary to integrate AI effectively across workflows.

3. Challenges of Scaling Adoption

Despite the successes, libraries face barriers to scaling AI adoption:

• Resource Inequities: Smaller or under-resourced institutions may struggle to keep pace with AI advancements.
• Workforce Preparedness: Polarization among staff, with varying levels of enthusiasm and skepticism, underscores the need for tailored training and consistent communication.
• Ethical Complexities: Libraries are grappling with issues of transparency, privacy, and responsible use, necessitating the development of robust ethical frameworks.

4. Early Adoption with an Eye on the Future

While some respondents noted that their institutions are still in the early stages of adoption, there is a clear sense of momentum. Libraries are strategically balancing short-term experiments with long-term vision, ensuring that AI adoption aligns with institutional goals and community needs.

Broader Implications

The findings from this poll underscore a critical inflection point for libraries. As AI tools become more integrated into workflows, libraries must:

• Advocate for equitable access to AI technologies to bridge resource gaps
• Lead campus-wide discussions on ethical AI use, positioning themselves as thought leaders
• Foster a culture of experimentation and learning to build confidence among staff and users

Libraries are uniquely positioned to navigate the opportunities and challenges of generative AI, blending technical innovation with their long-standing commitment to equity, ethics, and community engagement.

The findings from this third poll underscore the steady integration of generative AI in libraries. While optimism and adoption have increased, challenges such as ethical considerations, employee skepticism, and resource disparities persist. Libraries are making strides in professional development and collaboration, positioning themselves as leaders in navigating the opportunities and risks of AI technologies.

Continued research and dialogue are essential to address emerging challenges and ensure that libraries can harness the full potential of generative AI responsibly and effectively.

To stay engaged and receive the latest ARL updates, register via our website.