TECH TALK

Comment on this article

Algorithm Awareness

by Casandra Laskowski

Algorithms are everywhere. They are invisible gatekeepers and judges. They inform our decision-making, affect our search results, decide whether we get seen by future employees, target ads to us, and control the news we see. As ubiquitous as algorithms are in the modern world, librarians need to understand what they are and how they influence the world around us, both for our own sake and for our patrons.

In simplest terms, algorithms are computer code. They are a set of instructions for how a computer should handle data, solve a particular problem, or perform a specific task. These rules are often inaccessible and opaque, which prevents a full understanding of how they come to their final result. It is also important to note that algorithms are only helpful if they have data, which brings up issues of privacy that librarians continually face with new technologies. However, they are neither good nor evil. They are not altruistic or malicious. If you could attribute emotion to them, it would be indifference.

This indifference does not mean they never have an ill effect. They can be biased in ways that affect peoples’ liberty, their jobs, and more. Which is why we must be vigilant and informed about algorithms.

A basic understanding of algorithms will make librarians better resources for our patrons. We are adept at teaching information literacy. Google has an algorithm for deciding what websites to display to a particular search and Facebook has one to determine what you see on your news feed. While we cannot uncover their secret sauce, teaching patrons how to navigate searches and social media has been part of many librarians’ other duties as assigned. Similarly, we can teach a more general algorithmic literacy.

We can show our patrons when algorithms might be influencing their lives, so they can better adjust their behavior to get the desired result. Are they applying for a job? While many employers are automating parts of their hiring process, it can be frustrating to be on the other side when you do not realize you are interacting with a bot. Explaining how applicant tracking systems determine whether a resume gets seen better equips the patron to optimize their resume for computer review.

Additionally, a good foundational understanding of algorithms allows librarians to make informed software adoption decisions and head off potential problems before they take root. In one unfortunate example, a Florida library director was suspended for his part in a scheme to circumvent a flaw in a weeding algorithm—the removal of classic books from the collection. The software was apathetic to the quality of the books or the legacy. It did as instructed; it looked at circulation data. Any librarian with some collection development experience would understand that circulation information is insufficient on its own to inform weeding decisions.

It is unclear how involved any of the library personnel were involved in the decision to adopt the software. However, a few simple questions might have prevented this problem by poking the company on its algorithm. What criteria does it use in determining candidates for weeding? Is there a method to tag certain books (e.g., classics) so the algorithm gives them more weight than they might otherwise receive? The right questions might allow us to peer, superficially, into the algorithmic black box.

For example, there might be a great recommendation for software a library may want to employ. The right questions can ensure if it is a good fit for the organization. Does it require individual checkout data? Or can it make recommendations based on, for example, five book titles? Can patrons tag books in their record as anomalies (e.g., required reading not leisure)? The answers may cause a library to draft a detailed explanation that patrons must read before opting in. Alternatively, in the interest of patron privacy, a library might choose software for the reference librarian to use to assist with their recommendations.

Algorithms can be helpful tools for expedient answers and more accurate analysis. However, they can also be biased and faulty. By asking the right questions and maintaining human oversight in the process, we can improve library services while mitigating some of the adverse effects they might have in our libraries. As always, regular review must be a part of the process to catch those issues we did not predict.


Copyright 2018 by Casandra Laskowski.

About the author: Casandra Laskowski is a Reference Librarian and Lecturing Fellow at Duke Law. She received her J.D. from the University of Maryland School of Law, and her M.L.I.S. from the University of Arizona. Prior to pursuing her career as a law librarian, she worked as a geospatial analyst in the United States Army and served a fifteen-month tour of duty in Iraq. Her areas of interest include privacy, censorship, and the intersection of national security and individual liberty.