VOICES OF THE SEARCHERS
From Boolean to Bots
by Mary Ellen Bates
We librarians have always stood with one foot in the past and the other boldly striding toward new technologies. Back when I was in library school in the early 1980s, when Earth’s crust had not yet fully hardened, we had a great schism between the students who were focused on computer science and those focused on children’s literature. We each thought the other group was missing the point of getting a master’s in library science. It wasn’t pretty.
Fast-forward 50 years, and I am seeing a similar conversation. As I write this, I am waiting for a call from a local university professor who would like my thoughts on how she can teach her students to best use ChatGPT for their research. When she initially approached me, I demurred, saying that my advice to her students would be to use generative AI before and after the search, but to leave work that requires a human’s discernment, judgment, and intuition to humans. She persisted, so we will be having a conversation about AI literacy, the difference between a response that sounds plausible and the correct answer, and a chatbot’s inability to think outside the box to identify new insights. I will remind her of the value-added databases the university library subscribes to and why her students will be better prepared for life after graduation if they know how to evaluate information sources, identify bias, and spot gaps in what is known about a topic. And I will recommend that she introduce her students to Google’s NotebookLM and Claude projects—AI tools that enable the students to query the results of their research for deeper analysis.
As I give workshops and presentations about the role of information professionals in the new world of ubiquitous AI, I am struck by the need for info pros to lean into our core gift of being information whisperers—people who understand how information is organized and accessed, how and when their users are likely going to be seeking information, and what information technology is most appropriate for each information need. We have to become as proficient in using AI tools as we are with using professional databases and other resources for information discovery.
The more I reflect on how librarians and info pros can keep our professional skills updated and relevant, the more I see that the guiding principles of librarianship I learned decades ago still apply today:
- We have always focused on enabling critical thinking and teaching users how to evaluate information sources and tools. Previously, that took the form of teaching information literacy; today, we teach AI literacy (and AI skepticism).
- We understand how information is structured and how to access it most effectively. In the days of Boolean logic, we were the ones developing complex search strategies and arcane approaches; now we work as in-house AI prompt engineers
and ethical AI advocates.
- We are always bringing in new information discovery tools. Libraries drove the introduction of digital content to users and patrons back in the 1980s; today, we are being brought into project teams to support text and data mining initiatives and curate high-quality datasets.
- We continue to find new ways to enhance reference services. Discovery tools such as libguides and OPACs are being overtaken by custom-designed chatbots and language models comprising libraries’ special collections.
Whenever I read the weekly Department of Data column in The Washington Post, I am reminded of where we humans far surpass AI. As Andrew Van Dam, a data nerd and self-described “libraryphile,” describes it, he takes “overlooked and underappreciated data sets in search of hard truths and keen insights that will help make you smarter, better-looking and more fun at parties.” In a recent column, Vam Dam looked at how to use Google Trends to figure out what bugs people are trying to irradicate by season and region. By looking for the frequency of the phrase “how to kill [ants/fleas/fruit files],” he graphed out where we’re trying to kill which insects. I can skim the maps to see that carpenter ants bug people in Maine and boxelder bugs are annoying residents in North Dakota. Knowing an insect’s geographic range is one thing; seeing where people are irritated enough to Google how to kill them is another.
This kind of curiosity about what unexpected insights and correlations might be unearthed from a dataset is second nature for info pros but is beyond the capacity of most AI tools. The core skills of professional librarians remain as crucial and relevant in today’s digital age as they have ever been. |
