AI is a hot topic right now, especially since the launch of ChatGPT. However, it has been permeating our profession for a while, including in an area I have researched – bibliometrics. Prior to the kōrero, I shared one example of a pre-ChatGPT citation classification tool using AI in the conference presentation titled ‘Artificial intelligence (AI) and libraries: what is the fuss?’. However, ChatGPT feels like a watershed moment, a significant evolution in our understanding and experience of AI. Unsurprisingly, the kōrero session was oversubscribed, with people turned away due to lack of room. I hope this article will go some way to acknowledging the generosity of discussion on the day and allowing those who missed out a chance to also engage with the topic.
The goal of the kōrero was to give attendees the opportunity to broadly engage with the concept of AI and begin to think critically about both the opportunities and challenges it brings to the profession. The first task was for groups to come to a shared understanding of what AI means to them. After all, people conceptualise it in various ways. For some, it conjures terminator-like images of robots and intelligent machines with enhanced human capabilities. For others, neural networks and learning algorithms use our data to recognise patterns and make predictions or decisions.
At the kōrero, reoccurring descriptive themes included AI being a technology, a software, an algorithm and a tool. AI was described as a product by humans for humans. In other words, it is fundamentally a creation of human ingenuity programmed by humans to serve human needs and interests. Yet it was also described as a black box requiring a cautious approach. An external intelligence transcending that of humans on specialist tasks and domains, yet capable of stupid things like hallucinating citations, generating misinformation and plagiarism.
Moving on to the pros and cons of AI, attendees already had a broad knowledge of common problems or concerns. Some problems described were at the micro level. For instance, when ChatGPT returns a non-existent reference to a specific prompt. However, it seemed that the macro-level issues were the ones many attendees sought more direction on. These are issues around AI and data sovereignty, mis or dis-information generated by bad actors, lack of regulation, bias in training data, impact on intellectual property and privacy rights, lack of empathy, and potential job losses. Equally, attendees identified many positives to AI, both potential and realised. As a productivity tool, AI can support creative thinking, generate ideas, remove accessibility barriers, enhance outreach, and reduce cultural tax. Rather than making libraries redundant, it further demonstrates the value and relevance of libraries. However, to advance the positive side of AI, library management needs to be prepared to make space for professional development and experimentation.
The final part of the kōrero was a call to action to determine our possible next steps. Upskilling ourselves and educating our communities was an important next step for attendees. Take time to think, experiment, share, use, and critique AI tools and do it now as these continue to become more prevalent and improve. As one attendee expressed it, “Tuwhitia te nopo! Feel the AI and do it anyway”. The call was also made to seek more clarity from regulators, policymakers, individual institutions, and professional leaders on issues like AI and copyright, harvesting library data, intersection with traditional knowledge, and use as a research tool.
My reflection from the kōrero is that two streams of considerations are currently at the forefront of our profession regarding AI. One is a personal one focused on learning how to use specific tools, how to teach others to use them, and how to sort the wheat from the chaff in this rapidly changing environment. Professional development is central to this stream as it provides a means to meet those challenges. The other stream is at the macro level, asking the bigger questions about AI encompassing ethics, law, psychology, autonomy, and sovereignty. We are seeing more conference presentations, webinars, training courses, and other tools supporting our professional development with AI. However, the big questions are much further away from being answered. The rapid pace of AI advancement continues to outpace the development of policies and regulations to govern the emerging ethical, legal, and societal implications.
If this is a topic you are interested in, I have curated some resources of various types for you to explore:
- The best introductory video I have seen to address AI fluency specifically for librarians: https://vimeo.com/841055109/056df1c313
- My go-to article for definitions of AI by Andrew Cox and Suvodeep Mazumdar: Defining artificial intelligence for librarians: https://journals.sagepub.com/doi/full/10.1177/09610006221142029 [open access]
- Andrew Cox’s recent webinar presentation for LIANZA TEL SIG: Data, AI and Digital Transformation https://www.youtube.com/watch?v=lfi-QPvNLd8&t=2630s
- Lorcan Dempsey’s blog post: Generative AI and libraries: 7 contexts. You may wish to explore some of his other posts on the topic: https://www.lorcandempsey.net/generative-ai-and-libraries-7-contexts/
- My favourite library blogger: Aaron Tay is always testing the latest AI developments and providing an honest and accessible assessment: https://musingsaboutlibrarianship.blogspot.com/
- Join IFLA’s AI special interest group: https://www.ifla.org/units/ai/
- Explore the AI4LAM YouTube channel for a variety of examples and perspectives: https://www.youtube.com/@ai4lam120/videos.
“Hasta la vista, baby.”
Shiobhan Smith is the Associate University Librarian (Customer Experience) at Te Whare Wānanga o Otāgo the University of Otago. She has more than 17 years of experience working in academic libraries, including roles educating researchers on open access and supporting the use of OUR Archive, the University of Otago’s institutional repository.