A launch date has not yet been announced, but OverDrive has been recruiting librarian advisors from public, academic, and K–12 libraries to offer input and refine Readtelligence—an upcoming suite of tools for ebook selection and curation developed by the company using artificial intelligence (AI) and deep learning tools to analyze every ebook in the OverDrive Marketplace.

“I just want to ask librarians, ‘If you had a tool that could point at a catalog of titles and supplement the metadata, the bib record, the MARC, the BISAC, the LOC [subject headings], the genre…what else would you like to know?’” OverDrive President and CEO Steve Potash told LJ.

Potash described the new tools as “an R&D project that’s been a natural evolution of over 25 years of quality control and enhancing the management of digital book inventory,” adding that the development of Readtelligence had been driven by feedback from librarians. “I’ve been cataloging challenges we’ve heard from our [customer] librarians…. There’s been a variety of gaps in their ability to [evaluate] new books, new series, new graphic novels, manga, the list goes on and on.”

Readtelligence was first announced on August 5 during the “Your OverDrive Service of Tomorrow: Advancing the Science” presentation by Potash at the company’s biennial Digipalooza conference, held virtually this year. As LJ reported then, he explained that the new tools would give libraries the opportunity to use AI-driven analytics tools similar to those employed by major publishers and Amazon. Librarians could use the tools to highlight titles with specific themes or frequently mentioned locations or people to supplement purchasing decisions or curate themed collections from their library’s ebook holdings. The tools can also be used to generate and analyze metadata such as the average length of words, sentences, and chapters in OverDrive ebooks, facilitating curation by reading level or other criteria. In addition, new tools such as an AI-generated “emotion curve” will enable librarians to use entirely new metrics to compare fiction titles for purchase or curation.

“Of course, we have Lexile score, we have ATOS, we have grade leveling, genre, etc. But when I hear educators…they want more granular ability to create a bookshelf for an audience of students,” Potash said. “I went to my team and asked, ‘Are there any objective data points that we could take out of the book?’ And we started cataloging dozens of them.”

Offering a specific example of a problem that Readtelligence could help solve, Potash noted that while most major publishers offer metadata such as Lexile scores for new ebooks, many smaller trade publishers do not. When this information isn’t provided by the publisher, K–12 school librarians have no simple way to gauge the reading level of a group of new titles. “If a librarian wants to curate a summer reading list for a target audience of young boys high/low [high interest, low reading level] titles…imagine the ability for an educator or school librarian now to say, ‘How do I go into all of this graphic novel material and see if something might have this interest level and this level of text complexity?’” Potash said.

In addition, Potash said that the tools could be used to facilitate diversity audits. “Somebody could say, ‘Quickly scan all of the books I have,’ and they could even enter in, or check the box for all of the terms that they want to surface as part of a DEI audit. Right now, you can search by keyword in our catalog, but I’ve always said that’s not enough. Because [keyword search is] solely dependent on the metadata, the short description, the long description, the keywords put in by the supplier, the publisher, [and] the aggregator.”

Readtelligence will also bring full-text indexing to OverDrive Marketplace, which will enable librarians to conduct proximity searches and search the company’s entire catalog using Boolean operators. “There’s parallels to what we’re doing in academic library catalogs,” Potash noted. Drawing a comparison with full-text academic databases, he said, “You have the ability to see what’s inside on every page.”

During his Digipalooza presentation, Potash briefly noted that some K–12 educators and librarians have complained to him about explicit content in juvenile titles, and that Readtelligence tools could also be used to screen for such content. Using the virtual meeting’s live chat function, a few attendees expressed concern that these features could be used to block certain types of content from being considered for purchase. In the follow up interview with LJ, Potash said that Readtelligence was created to give librarians and curators more granular information and help facilitate buying and curation choices, not to take choices out of their hands. Implications of blocking or censoring content are “totally outside of the scope of the project,” he said, later emphasizing that “it doesn’t replace curation, and it doesn’t replace selection.”