Artificial Intelligence and Libraries

This report concludes with a series of recommendations of actions that the AU Library can take to promote equitable access of AI tools to the entire campus community. These recommendations reflect the challenges and opportunities equitable access to AI poses and the values of our institution and profession. They include:

• Recommendation 1: Removing Cost Barriers
• Recommendation 2: Equity of Understanding
• Recommendation 3: Assessing for Equity in AI Tools
• Recommendation 4: Building Equity Through Transparency
• Recommendation 5: AI Tool Guide

The path toward integrating AI tools and methodologies into the AU Library's offerings is not just about embracing new technologies; it is about ensuring these advancements serve everyone equitably. Our recommendations are rooted in the belief that access to these powerful tools should not be a privilege but a shared academic resource that uplifts and supports all members of our community. Moving forward, our commitment to equity should guide the implementation of these technologies. We aim to create an environment where every student and faculty and staff member has the resources they need to succeed, irrespective of their starting point. By doing so, we not only enhance our library's offerings but also reinforce our university's commitment to fairness, inclusion, and equal opportunity in the digital age.

Library users cannot take advantage of tools that they lack access to. Knowing that the most powerful and sophisticated AI tools are paywalled, the library can level the playing field where appropriate with the following:

• Recognizing the AU Library can only license tools that support institutional subscriptions and has budgetary limitations, acquire AI tools that support the AU Library’s mission and strategic plan, and that align with the broader scholarly initiatives of American University.
• The Library Leadership Team needs to determine the appropriate body to make determinations about pursuing institutional licenses to AI tools and whether it lives within the current structure or if we need to develop a different structure. Similarly, the Library Leadership Team needs to determine the appropriate budgetary stream for such purchases.
• Develop criteria to distinguish between AI tools that are best provided and supported by the AU Library from AI tools that should be housed elsewhere or live at the enterprise level. For AI tools outside the scope of the AU Library, the AU Library should collaborate with campus partners and advocate for library users when appropriate. In doing so, the campus partners should strive to avoid user confusion regarding which AU unit provides what kind of access to, and support for, which AI tools.
• Task the appropriate body/individuals within the AU Library to assess current software acquisitions policies and procedures to reflect a commitment to equitable access of AI tools.
• Redouble efforts to secure additional funding for the Inclusive Technology Initiative, thus providing all AU students with adequate hardware to access AI tools.
• Recognize that different academic disciplines and schools will incorporate AI at varying rates and to address unique needs. Ensure that acquisition policies and procedures are flexible enough to account for these differences.

Scholars cannot capitalize on AI tools that they do not comprehend. Recognizing that equitable access to AI is meaningless without an understanding of how to engage with that AI, the following is recommended:

• Actively engage with the recommendations and proposals set forth in the sections on Information Literacy and Privacy, Copyright, and Intellectual Property. Those proposals in conjunction with current information literacy and scholarly communication practices and instruction can provide the campus community with the requisite skills and knowledge to equitably access AI.
• Acknowledge and address head-on the problems with AI that decrease trust – bias, deepfakes, misinformation – and concern our community. Develop a feedback mechanism for users’ concerns to be incorporated into the AU Library’s decision-making process for AI adoption.

Provide more training on AI, its relationship to the library, and any library-provided AI tools to the entire library team to meet user needs. Every user deserves the highest quality service and assistance wherever and whenever they engage with our resources.

To ensure the equitable integration of AI tools within library services, it is essential that Library Leadership evaluate current policies and develop a comprehensive framework for assessing AI technologies prior to licensing or recommendation to users. This involves adapting software acquisition policies to include AI tools, actively discussing the tools' limitations alongside their benefits, and instituting a thorough evaluation process for AI features in existing and new library products.

• Conduct a comprehensive review of existing policies to identify gaps in AI tool assessment and develop specific criteria for evaluating the inclusivity, accessibility, and representation within AI technologies. Given the speed of AI technology, this review should be periodic.
• Ensure the guidance on AI tool acquisition aligns with the library's mission and the diverse needs of its user community, focusing on enhancing accessibility and inclusivity.
• Establish transparent procedures for the evaluation of AI tools, including discussions on their limitations and biases. Engage with both vendors and library users in this process to ensure a broad understanding and mitigation of potential issues.

While not a cure-all, transparency by libraries can reassure library users as they navigate the upheaval of AI and address numerous dimensions of equity of access. Libraries should actively and openly communicate about the opportunities and challenges of AI, particularly as it relates to the library and equity of access and seek regular feedback from the diverse community of library users.

• Develop a communication strategy that describes the AU Library’s evolving approach to AI, how the library evaluates opportunities and risks posed by AI, and clearly explains the resources and services we provide.
• Create an accessible and easily discoverable online space to share the Library’s approach to AI, and include lessons learned as our approach to AI changes with time.
• Engage with students and faculty regularly to understand their changing needs for and perceptions of AI as the technology matures. This can be done though surveys and assessments, usability studies, and listening sessions.

As central to the university mission and critical to information access and creation, the academic library is a place that should explore, assess, and develop expertise in free and viable AI tools. Limited budgets and unlimited needs will prevent the library from acquiring every AI tool that enters the market or is requested by a user. These recommendations are intended to close that gap presented by limited resources and connect users with AI tools they may not discover themselves.

• Identify and task the appropriate library body with developing a guide of library-provided AI tools and methodologies. The guide should include information on how the tools/methodologies were selected by the Library, information on the underlying data used to develop the tool/methodology, and practical applications.
• Identify and task the appropriate library body with developing a guide of free AI tools and methodologies. This guide should include information on how the tools/methodologies were assessed by the Library, information on the underlying data used to develop the tool/methodology, and practical applications.
• Promote American University’s AI-based research and learning in the Library. Engage with multiple modes of communication – academic posters, events, and exhibits – and help library users connect this promoted research and learning with their own academic endeavors.