Artificial intelligence (AI) is a rapidly evolving technology that can potentially transform librarianship in various ways. AI can be defined as “the ability of machines to do things that people would say require intelligence” (Jackson, 1985). AI can enable new capabilities to address users’ information needs, such as providing not just information but deep intelligence and offering “Insight As A Service (IAAS)” (Springer Nature, 2019). While AI can also enhance and augment many library services and workflows, such as content indexing, classification, search and discovery, data analysis and visualization, chatbots, and virtual assistants, AI, however, also poses some challenges and ethical issues for librarians, such as privacy, bias, accountability, transparency, and trust.
As librarians, we need to be aware of the impact of AI on our profession and our users and actively participate in designing and evaluating AI-based tools and applications. We also need to update our skills and knowledge to keep up with the latest trends and developments in AI and related technologies.
We will explore some of the ways in which AI would change the role of libraries in these three main areas:
- Content Indexing
- Virtual Assistants and
- Data Visualisation
We will also look into opportunities that we, as a librarian, can leverage AI to improve library services.
Content indexing is the process of assigning keywords, subject classification or metadata to documents or any library resources to facilitate content discovery for retrieval and use. This has always been done manually, and it proves to be a tedious task even for experienced librarians. The process typically involves analysing the text of the content to identify keywords, phrases, and other metadata that can be used to describe the content and facilitate retrieval. Naturally, this is a vital process of any library information system, not limited to search engines, content management systems, and digital assets management systems.
Libraries can leverage the speed of AI to automate and improve this process by utilising natural language processing (NLP) and machine learning (ML) capabilities to analyse digital collections and identify relevant topics and thus assign suitable metadata to improve search and discovery. We look at how Springer Nature employs AI to index its vast collection of eBooks and provide semantic search capabilities (Springer Nature, 2019).
We can tap into this capability and benefit from AI-powered content indexing by accessing more comprehensive and accurate metadata in our collections, which at the end of the day, shall improve search results with precision. Similarly, ML may group or cluster documents of similar concepts as a result of enhanced classification schemes, taxonomy and ontologies.
Virtual Assistants (VA)
Virtual assistants or digital assistants are applications that use natural language understanding (NLU) and speech recognition to interact with users via voice or text (Read: 26 Actually Useful Things You Can Do with Siri). VAs may perform various tasks for users, such as answering questions, providing information, setting reminders, playing music or controlling smart devices. Some of the most popular VAs are Google Assistant, Siri, Alexa and Cortana.
VAs have also entered libraries in the form of chatbots that can handle directional questions on a library website, alert when a book is due, point a user to relevant library resources or answer simple informational requests (Hervieux & Wheatley, 2020). In the future, VAs may be able to provide more advanced services for our users, such as recommending books or articles based on user preferences or context, summarizing key points or arguments from a document or dataset, or generating citations or bibliographies. We can use VAs to enhance our user experience and satisfaction and reduce their workload for repetitive or routine tasks. We can also collaborate with VAs developers to ensure that they are reliable, accurate and ethical.
Generally, data visualization is the process of presenting data in graphical or pictorial forms to communicate information effectively and efficiently. This can help or aid users to better understand complex data sets, discover patterns or trends, compare variables or categories, or tell stories with data.
Data visualisation tools can use AI to analyze data and generate visualisations automatically or semi-automatically based on user input or preferences. For example, SciGraph Explorer is a data visualization tool that uses AI to map connections among concepts, researchers and institutions based on Springer Nature’s publications (Springer Nature, 2019). We can use data visualization tools to provide IAAS for our users by helping them find unexpected connections or insights from existing data sources. We can also use data visualization tools to showcase our own collections or services in an engaging and interactive way.
AI is changing libraries in many ways by offering new possibilities for improving library services and workflows. We need to embrace AI as an opportunity rather than a threat by learning about its potential benefits and challenges for our profession.
- Jackson, P. (1985). In Introduction to artificial intelligence. essay, Dover Publications.
- The impact of Artificial Intelligence on librarian services | For Librarians | Springer Nature. (2019, July 21). The Impact of Artificial Intelligence on Librarian Services | for Librarians | Springer Nature. https://www.springernature.com/gp/librarians/news-events/all-news-articles/ebooks/the-impact-of-artificial-intelligence-on-librarian-services/16874432
- Wheatley, A., & Hervieux, S. (2020, February 6). Artificial intelligence in academic libraries: An environmental scan. Information Services & Use, 39(4), 347–356. https://doi.org/10.3233/isu-190065
- IFLA Trend Report — Advances in Artificial Intelligence. (n.d.). IFLA Trend Report — Advances in Artificial Intelligence. https://trends.ifla.org/literature-review/advances-in-artificial-intelligence