FEATURE
Learning About Search at the CIL Conference
by Marydee Ojala
The Computers in Libraries conference (CIL; computersinlibraries.infotoday.com), held during the height of cherry blossom time in the Washington, D.C., area, covered a lot of ground, as reflected by its theme of Libraries Adapting to Change. From the search perspective, nothing has changed more in the 40 years the conference has been going on than search technology. Computers were a bit of a novelty in the mid-1980s, not something that people routinely used. Phones were anchored in place and hardwired, not the smartphones (which are really mini-computers) we carry around today. In fact, the original name of CIL was Small Computers in Libraries
(SCIL), and “microcomputers” were a hot topic.
Adapting to change has been a mainstay of librarianship for decades, if not centuries. For a humorous take on this, look at Norwegian television’s video (youtube.com/watch?v=pQHX-SjgQvQ) of the help desk guy explaining to a medieval monk how to transition from using scrolls to interacting with a book. Changes in search have been happening at a vastly more rapid pace in recent years. It seems that as soon as you learn about a new development, it’s obsolete and you have to learn about an even newer develop ment. It’s disorienting and head-spinning.
WEBSEARCH UNIVERSITY PRESENTS SEARCHERS ACADEMY
Librarians who opted to attend the preconference WebSearch University presents Searchers Academy workshop heard Arthur Weiss, AWARE (a U.K. consultancy specializing in competitive and marketing intelligence, research, and training), speak about many striking changes in search. He pointed out that search engines such as Google are jumping on the AI bandwagon with great alacrity. Many Google search results now put AI Overviews at the top, with the blue links underneath. Google Lens lets you use a circle to narrow your search to part of an image. Keyword search is not completely dead, but could be in the near future, Weiss speculated.
Bing is bringing its Copilot AI capabilities to optimize search results and their appearance, using generative search derived from OpenAI’s collaboration with Microsoft. Copilot encourages natu ral language queries. Microsoft Search in Bing (for businesses and colleges) was ”retired” at the end of March but is still accessible via m365.cloud.microsoft and SharePoint Online. DuckDuckGo’s foray into AI can be found at Duck.ai or using the !ai and !chat bang shortcuts. It still emphasizes privacy as a central tenet but does allow for localization if desired. Considering pure AI search ing, Weiss briefly discussed Perplexity, You.com, and ChatGPT.
Weiss commented that fuzzy match searching is now relatively easy, keywords have has become much less important, and natural language (semantic) searching is now the normal way to search. He wondered if generative AI (gen AI) is the real Google killer. He concluded with some expectations for the future of search—a further emphasis on privacy and personalization, more AI tools and AI plugins, greater emphasis on monetization of AI, more dis- information appearing, and an overall shake-out of AI tools due to mergers and acquisitions.
My contribution about savvy searching included some decision points about choosing wisely about where to search— traditional library databases, general web search engines, specific websites, or new gen AI tools.
SEARCH & DISCOVERY
The Seach & Discovery track asked if search was evolving. It seems more appropriate to say that it is evolving and then ask what librarians can do to adapt. Mary Ellen Bates, Bates Information Services, provided searcher tips. Her take on gen AI tools was to use Gemini, Copilot, and ChatGPT for current information; Consensus, Perplexity, and Elicit for peer-reviewed sources; and Claude and ChatGPT for advice or analysis. She recommended Brave (search.brave.com) as a search engine and Tara Calishain’s work creating new and exciting search tools (researchbuzz.me).
Noting that search in the AI context is all about conversations, Bates had several suggestions for getting the conversation started, including instructing it to say so if it doesn’t have enough information to answer the question, to ask questions to clarify the request, to combine ideas from the answers, to expand on a particular idea, and to give specific examples. Another suggestion was to ask the bot to improve the prompt.
Reasoning models are on the horizon. Bates regards them as a threat since it can all too easily lead to people within our institutions deciding that librarians can be replaced by searchbots. She recommends getting familiar now with reasoning models and putting them through their paces to position yourself as a “searchbot sherpa,” guiding patrons through their search journey.
ASCENDANCE OF GEN AI
Gen AI affects many aspects of our library life; some are peripheral to search, others, integral to it. Exploring AI had an entire track devoted to the technology. Another track concentrated on AI Insights & Challenges. Even in tracks not specifically devoted to AI, sessions about AI issues appeared. AI literacy, for example, was in the track about Learning New Models of Librarianship.
Amy Affelt, Compass Lexecon, traced AI from digital assistants to gen AI and then to agentic AI. Perplexity can now search, write, and create a report. Google’s Project Jarvis is poised to take over your browser to gather research, purchase a product, or book a flight. Gemini’s Deep Research can search for a topic, prepare a report of key findings with links to sources, and upload the report as a Google Doc. NotebookLM can create podcasts. Like Bates, Affelt thinks that it’s incumbent upon information professionals to keep abreast of AI technologies and be the go-to expert.
A very practical approach was provided by Christine Carmichael, Creighton University, as she detailed specific projects the library spearheaded, including an introduction to AI for students, faculty workshops about integrating AI competencies into the core curriculum, and piloting three AI research assistant tools. She said it really helps that the library director’s passion is AI and ethics. Ethical con siderations included copyright (still not resolved by the courts), labor (AI taking people’s jobs), GIGO (How bad can the internet training be? Probably very.), lack of transparency, and bias.
AI literacy was addressed by Anne-Inger Hellekjær, Oslo Metropolitan University Library, and Jesper Solheim Johansen, Keenious, who reported on a study they did to determine how library school students view AI, how AI usage has changed, and whether new students perceive themselves as more or less optimistic than previous students. The pair determined that greater knowledge does not necessarily lead to optimism, but using chatbots tended to increase learning and motivation. However, a need for guidelines about us ing AI also exists, as AI is inevitable. Thus, libraries will find the technology particularly significant, they found.
Jeff Wisniewski, University of Pittsburgh Library System, said that creating an AI-literate library starts with strategy. Assess community needs for AI literacy and determine your learning objectives, such as basic AI concepts and definitions, knowledge of gen AI models, skill in using AI tools, assessing output of the tools, and understanding ethical implications. Note legal and policy aspects, along with any library-specific considerations. He recommended piloting small AI projects and assessing impact and developing policies for responsible AI use.
AI INSIGHTS AND CHALLENGES
AI is not without challenges for the library community. Matt Clemons, University of Arizona, reported on his 3-month internship project, calling his talk “Anything You Can Do, AI Can’t Always Do Better!” He was charged with taking a somewhat limited dataset from HeinOnline—article titles from AAAL Spectrum and Law Library Journal—and analyzing them to understand the evolution of law librarianship over time. After some necessary data cleansing, he created topic models, vectorized titles, applied clustering algorithms, and analyzed the results. Clustering produced some incoherent groupings, such as “Microforms on Legal Subjects” with “About Microaggressions” and “The Law Library Mission Statement” with “Sourcebook on German Law.” He concluded that humans can succeed where AI fails.
Brian Pichman, Evolve Project, reminded the audience that technology is inherently neutral, but can be used for good or bad things. His favorite tools include Rytr.me to generate content, supermachine to generate art, and Steve AI to create videos. Nicole Hargrove, UNC University Libraries, wrote a student guide to gen AI and has a public service model to empower the campus community regarding gen AI. Her internal-facing efforts involved personas for staff—productivity booster for all staff, customer service supercharger for circulation staff, workflow guru for tech services, and research and instruction specialist for liaisons.
Although CIL speakers explored topics other than search, that’s where I concentrated my learning activities. Library technology continues to display incredible innovation—and not just where AI is concerned. The luscious pink of the cherry blossoms outside the conference venue provided a contrast to the intensity of the presentations inside, serving as a reminder that sensory experiences can provide a welcome balance to technology. Although gen AI can produce equally lovely images, being out among the real thing brings a different type of appreciation. If library technologies don’t enhance user experiences and help build community, they will be as fleeting as those cherry tree blossoms.
|
