Feeling feeble with your single human-sized brain? Want to tap into computational powers hitherto unknown to scribbling mankind?
You’re in luck: UK Libraries is introducing a new AI-powered “research assistant” into the search experience.
Developed by Ex Libris, the Primo Research Assistant tool allows users to explore academic content by posing questions in natural language. The tool searches nearly six billion records in Ex Libris’s Central Discovery Index to find the five “most relevant” abstracts, which it summarizes in an answer of about 200-400 words.
The answers include in-line references that allow users to click through to find the original source being summarized by the model. All five sources are listed below the answer, and users can click each source to find its record in InfoKat, with an option to view more results.
“The Research Assistant provides a new way to enter into our collections,” said Library Discovery Systems Coordinator Daniel Naas. “The normal search experience will not change. This is just a new feature that exists alongside the normal search experience that can help users uncover very specific sources that might not surface as quickly through a traditional search.”
Users can access the Research Assistant through InfoKat via a link on the top-of-page navigation bar. To use the tool, sign in with your linkblue ID.
Trained exclusively on scholarly content – and not the linguistic and informational wild west of the whole world wide web – the Research Assistant provides answers that are far more trustworthy and reliable than other Large Language Models.
“Interestingly, the tool won’t necessarily pick the same five abstracts each time it is asked a question,” said Naas. “So if you’re not quite happy with a search, you can ask the same question and you might get slightly different results or different sources.” There’s even a “Try Again” button at the bottom of the page to prompt the tool to “chew” on different sources.
For all of its capacities and (hopefully beneficent) power, there are still a number of limitations.
“For one,” said Naas, “it doesn’t understand resource types very well, so if you’re asking specifically for peer-reviewed sources or for articles from a specific journal, for example, it won’t be able to narrow down your search based on those kinds of parameters.” The same goes for creation date or time period, he added. “However, in the next feature release in early February, we’ll be getting a resource filter and recency filter, so improvements are coming soon!”
“The tool searches through a massive record set, regardless of whether we have full-text access to the content. Fortunately for us, given the breadth of our electronic resources, this isn’t a huge issue. But it is possible that some of the summarized sources could be from outside our collections, which users would need to request for pick up or delivery if they wanted to see the full text,” said Naas.
The Research Assistant also won’t search digitized content from UK Libraries’ local collections, like the Kentucky Digital Newspapers or items in ExploreUK. In addition, Ex Libris is still integrating record sets into the program, so the Research Assistant cannot currently search for materials from Elsevier or JSTOR, though it may be able to do so in the future.
“Lastly, it’s really optimized for natural language queries and for very quickly retrieving highly granular records,” said Naas. “A search for, say, ‘music history’ will give you an overview of the concept, but you’ll find the tool is much more useful for more specific questions, like, “How did American pop music evolve during the 20th century?’”
The tool is still in its Beta phase, with tweaks and upgrades expected in the near future. The Research Assistant does not share user data with third parties, and by engaging with the tool users are not “training” it.
Still, as it works out its kinks – and makes its plans to take over the world – give it a whirl and see what our future overlords have to say about the things that matter most to you. This human being with ten whole fingers and toes wondered:
Stay sharp, Cats!
