Artificial Intelligence applications for libraries

Artificial Intelligence (AI) pervades our everyday lives, whether we realize it or not. AI technologies power our digital personal assistants, show us traffic flows on our roads, monitor our health and diagnose diseases, guide our product selections when we shop online, and give financial advice. Libraries also benefit from AI technologies and AI applications are increasing.

Interest in finding innovative uses for artificial intelligence (AI) in libraries is growing internationally. Two online meetings the week of 6 December 2021 emphasized the importance—some might say inevitability—of AI technologies affecting library services.

IFLA Contemplates AI SIG

The first meeting, on 6 December, investigated the feasibility of creating a Special Interest Group (SIG) within IFLA for AI, under the auspices of the Information Technology (IT) Section. Organized by May Chang, University of Cincinnati and past chair of the IFLA IT Section, the need for a new SIG was triggered by the recognition that AI applications are already present in libraries in the guise of chatbots and robots, embedded in library systems, and used for automated indexing and classification. They have a future in content discovery and research support, front-facing customer service, smart buildings, and traditional automation integrated with robotics.

The meeting began with three presentations on the current state of AI in libraries, by Edmund Balnaves, CEO and Founder of Prosentient Systems and Chair of the IT Section, Juja Chakarova, Max Planck Institute Luxemburg, and Lynn Kleinveldt, Cape Peninsula University of Technology. Balnaves noted that AI in libraries had its origins in open source but the possibility of algorithmic malfeasance exists, as has become more apparent recently. The future may see contactless libraries. Chakarova joked that AI in libraries could stand for Already In libraries and wondered about delegating library functions to machines: Just think about indexing, search, translation, and image classification. Digital literacy was a concern of Kleinveldt. Information literacy has expanded into computational literacy. We now have smart learning environments and need to consider integrating coding and other aspects of data science into the curriculum.

Support for the formation of the SIG was overwhelming. Following its approval by IFLA, the next steps are to hold a business meeting to nominate a Convenor and seek volunteers to serve in roles including Secretary and Communications Coordinator.

The IDEA Institute Project

The second meeting, on 8 December, was presented by Naseej Academy and moderated by Abdallah H. Metwally, Naseej Academy General Supervisor and full professor at Department of Libraries, Documents and Information Technology, Cairo University, on the topic of “Artificial Intelligence (AI) in Libraries: Training for Innovation”. Speakers were Dania Bilal, Principal Investigator and co-developer, IDEA Institute on Artificial Intelligence, and Professor at the School of Information Sciences at the University of Tennessee in Knoxville and Clara M. Chu, Director and Mortenson Distinguished Professor, Mortenson Center for International Library Programs, University of Illinois at Urbana-Champaign. Bilal has also been instrumental in forming a Middle East regional chapter of ASIS&T. A video of the presentations is on YouTube.

Bilal began by explaining that AI could mean Assisted, Augmented, or Autonomous Intelligence, each of which has different characteristics. Thus far, the primary uses of AI technologies in libraries have revolved around chatbots and robots with a tech-centred emphasis on data extraction and linked data. Children, she said, engage more easily with robots than do older people.

Bizzy, an AI-powered chatbot at the University of Oklahoma, has an impressive 88% accuracy rate., a Colorado-based company, invites you to “Bot-ify Your Campus” with Ivy, its AI self-service chatbots designed for colleges and universities. At the University of Tennessee, the robot is called Tammy Pepper Knox, “but you can call me Tammy”. Libby is the robot at the University of Pretoria.

Chu gave five reasons why adoption of AI has been slow in libraries: A lack of expertise and training; social issues around algorithmic bias, privacy, and safety; challenges over fears of job loss and changes in job responsibilities; the evolving nature of AI; and financial constraints. The IDEA Institute on Artificial Intelligence   focusses on Innovation, Disruption, Enquiry, and Access to provide professional training to librarians and other information professionals. Understanding the challenges and opportunities in implementing AI in libraries is based on the real world experiences of the first cohort of Institute Fellows. Issues surrounding equity, diversity, inclusion and accessibility are also on the agenda, as are ethical questions.

She used the example of University of Illinois Engineering Librarian William Mischo’s “Bento Box Discovery Model” to demonstrate improvements in finding articles, items in the library collection and subject suggestions. She also mentioned Newspaper Navigator from Library of Congress Labs,  Wikidatabot's use of linked data and Google Cloud’s Dialogflow conversational AI that extracts data from Google.

Incorporating AI technologies into libraries has exciting possibilities both for librarians and library users. We at Internet Librarian International are fascinated by how libraries will evolve in light of developments in AI technologies, putting them to use to enhance library services and reach out to a wider population.