Lightning Round: AI Tools Secret Sauce; Developing Ethical AI Use Policies; Ethical AI in Libraries

In Search of the Secret Sauce: Questioning the Quality and Cost of AI-powered Research Tools
Tessa Withorn, Science Librarian, University of Louisville

AI-powered research tools are increasingly used by students and researchers, but are they fully aware of their limitations? Tools such as Elicit, Consensus, Scite.ai, Research Rabbit, Semantic Scholar and others are attracting researchers for their perceived ease of use, relevance, and ability to synthesize research. Although most tools have free versions with limited features, the increased commercialization of AI research tools highlights the information privilege of researchers with more funding and presents barriers for researchers who cannot afford them, which may disproportionately affect researchers from marginalized groups. In this commercialized environment, startups are also less likely to provide information about their indexing and algorithms, considering them the “secret sauce” that gives them an edge over their competitors. This presentation will give a few tips on how librarians can discuss the strengths and weaknesses of AI tools with researchers to best support them in conducting comprehensive literature reviews.

=====
Developing Policies for the Ethical Use of Artificial Intelligence in Higher Education and Libraries
April Sheppard, Assistant Library Director, Arkansas State University
Matthew Mayton, Archivist, Arkansas State University

The recent developments of generative artificial intelligence (AI) tools holds the potential to enhance all aspects of the teaching and learning process while also impacting the administrative and day-to-day operations of higher education institutions and libraries. In addition, the demand for college graduates to have high technical skills, including AI skills, are expected to rise exponentially. To ensure that our students are best prepared for a future-proof workforce, higher education institutions and academic libraries cannot shun AI. Instead, these organizations must balance the need to maximize these rapidly evolving technologies while also conscientiously addressing the ethical concerns they bring. In order to ensure the responsible and fair utilization of these tools, institutions and libraries must develop policies to encourage their ethical use.

This presentation will provide sample artificial intelligence policy language from various higher education institutions and academic libraries. Topics covered will include the acceptable use of AI in the classroom, the role of faculty in making AI-related decisions, syllabus statements, AI use and detection, AI literacy, and library policies regarding AI. Participants will be able to compare and contrast policies to help them develop their own policies that work for their unique organization.

=====
Ethical AI in the Library
Dr. Mojca Rupar Korosec, Library Councillor, National & University Library of Slovenia

There must be a first look at data and the ethical aspect of its use as a key element for work and to what data refers. Data ethics builds on the foundations provided by information ethics and assesses data practices. An important question is raised what are the implications of data literacy for ethical librarianship?

Libraries can provide the ability to think critically about data in different contexts and to reflect on the impact of different approaches to data and information. The AI Literacy Imperative is the assumption that today we have a critical need to understand and be able to use the key aspects of AI literacy.

AI can bring many benefits to libraries, improving efficiency, search and access to resources by supporting research activities. When implementing AI in libraries, it is crucial to take care of ethical and security aspects, ensure data quality, and take into account user needs and requirements.

In this excellent place for the new role of libraries we need to be very careful because generative tools don't just need data - they drive it. Data can be biased and unreliable, which comes not only from the data itself but also from the design of algorithmic systems and how they are implemented (and who implements them). Ensuring that data will be key to helping users to be aware of how much they can trust the data and the technologies and techniques they use.