OverDrive held its eighth Digipalooza virtually via the Accelevents platform Aug. 3–5, 2021, and it ended with quite a bang. It was opened by OverDrive founder and CEO Steve Potash, who promised big things to come over the course of the 3-day event. Attendees had opportunities to hear sessions, participate in breakouts, do exercises, network, and visit the virtual exhibition hall. There were raffles, prizes, and trivia galore, with each day culminating in a $1,000 prize for library collection development. The digital venue was easy to navigate, and although online conferences can never really approximate what the real-life thing is like, the setup was pleasant, user-friendly, and one of the best I have experienced during the pandemic.
The first day featured a keynote by Librarian of Congress Carla Hayden and sessions on digital access for all and the future of listening. Day 2 saw Dominique Raccah (publisher and CEO of Sourcebooks) give the keynote, along with sessions from authors Andy Weir and Abby Jimenez and information on data and intelligence from library catalogs. Day 3 was where things got the most interesting. Although the final day’s keynote was given by academic and activist Anita Hill, and OverDrive announced the sunset of the OverDrive app in 2022, the real fireworks took place in the clash between a session on patron privacy and one on the future of OverDrive.
“Protecting Reader Privacy and Navigating Regulations” was presented by Juliana Cotto (policy counsel for youth and education privacy at the Future of Privacy Forum), Erin Berman (division director of the Learning Group at Alameda County Library and chairperson of the ALA Intellectual Freedom Committee’s Privacy Subcommittee), Michael Dodes (library coordinator for the Queens Office of School Library Service at the New York City Department of Education), and Erica Lazzaro (EVP for publisher services and general counsel at OverDrive). Lazzaro posed questions to the panelists about data privacy and how public, school, and other libraries should handle patron information.
Berman’s expertise in intellectual freedom was a standout. Her explanation of data minimization, or the principle that organizations and vendors should only collect essential information instead of everything they are able to, was one of the most salient points made. Just because you can collect data does not mean you should, and as libraries, we should be asking ourselves what data we truly need, why we need it, and how long we should be holding onto it before destroying it—and the answer is not in perpetuity.
All of the panelists agreed that following federal and state laws to protect patron data is key but not enough. Librarians should be also following high standards of best practices, going beyond the low bar of what the law requires. Dodes stated that the best protection of intellectual freedom is not collecting unnecessary data at all, because then it cannot be used for ill intentions. Simple acceptance or use of a tool does not equal informed consent. Berman wrote in the comments, “The lack of informed consent out there with these analytics products is frightening.” Little did we know that one of the most frightening analytics products was yet to be unveiled by OverDrive itself.
THE FUTURE OF OVERDRIVE
Potash unleashed a controversial data-based beast in his “Crystal Ball Report,” also called “Your OverDrive Service of Tomorrow.” Each Digipalooza features this session, giving OverDrive the chance to promote some of its upcoming features and cutting-edge technologies. OverDrive intends to integrate voice, facial, and context authentication in the service, but was not specific about how it intends to do so. Attendees in the comments expressed concern about biased algorithms and how facial recognition technology would pose problems for privacy in libraries. OverDrive also announced the expansion of a Patron Interest Dashboard, which uses data from patron interactions on the Libby e-reader app to distill users’ actions into insights, such as what patrons are searching for and not finding. But that was only the beginning.
The real bombshell of the session was the unveiling of “Readtelligence," billed as a collection of deep learning insights on individual collection items. Potash explained how dynamic insights would be created for each individual title to support its bibliographic record. These insights would include information about the characters, words, images, voices, tone, and even emotion of a book. Content awareness tools (CAT) would tell librarians and patrons what attributes are attached to each item, according to the AI algorithm. CAT would score items before patrons even read them, allegedly informing patrons whether the item houses any “objectionable material,” as Potash called it.
This announcement immediately perked the ears of the participants. Libraries are supposed to protect intellectual freedom and the freedom to read. Yet our primary e-resource vendor is looking to be the final arbiter of what content is appropriate and what the emotional tenor of a book is on behalf of millions of diverse patrons. The line between library mission and vendor profit suddenly became stark.
Labels have always been a contentious issue in libraries, with some librarians viewing them as censorship and others believing they simply aid in selection. If taking the former view, it can be easily extrapolated how content tags and attributes can be weaponized by parties looking to punish readers of specific content. Reading is no longer anywhere your imagination can take you; with OverDrive’s Readtelligence, you already know where you’re going, and someone can stop you from getting there if they want to.
Session attendees were frank with their opinions in the comments. One participant mentioned that the idea was “super fraught,” while another said, “Dictators would love this level of control over what their people read.” Another warned, “Professionals better be on the same page about censorship, or ‘content awareness tools’ will enable us to make decisions we shouldn’t make for our readers in the public library space.” A final comment said exactly what I was thinking the whole time: “Didn’t we just attend a meeting about be[ing] skeptical of vendors using patron data?” The intellectual whiplash between sessions was truly staggering. Should we be collecting as little data as possible about patrons in order to protect them and ensure intellectual freedom? Or should we acquiesce to OverDrive’s artificially intelligent algorithm that will know us better than we know ourselves before we even tap “borrow”?
The future of Big Data is here. And libraries’ biggest vendor has the collection bucket ready.