Chatbots: Automating Reference in Public Libraries

Chatbots: Automating Reference in Public Libraries

Michele McNeal (Akron-Summit County Public Library, USA) and David Newyear (Lakeland Community College, USA)
Copyright: © 2013 |Pages: 14
DOI: 10.4018/978-1-4666-3938-6.ch006
OnDemand PDF Download:
No Current Special Offers


The authors discuss their experience with using artificial intelligence and chatbots to enhance their existing web sites and information services in public library settings. The chapter describes their budget driven motivations for embarking on this project and outlines the development and implementation of the bots in their library settings. They show how the bots are positioned to enhance existing services and describe the various reactions to the bots from their patron base, and staff. Different implementations of the bots are highlighted (text only, animated talking avatar, mobile site, desktop help icon) as well as the differing levels of complexity of these different implementations. They address the oft posed question “Does AI spell the end of Reference?” and describe the InfoTabby code sharing project.
Chapter Preview

Chatbots To Extend Library Information Services


Our journey toward the implementation of AI chatbots began at the Mentor Public Library (MPL), following then Ohio Governor Ted Strickland’s June 19, 2009 proposal to cut state funding for public libraries (Office of the Governor, State of Ohio, 2009). At that time the proposal included cuts which would have removed about 50 percent of the Ohio’s funding for public libraries between 2010 and 2012. This was on top of a 20 percent reduction already faced by Ohio’s public libraries due to decline in the state’s General Revenue Fund (Sun News, 2009). As the Sun News (2009) reported, “Nearly 70 percent of the state's more than 250 public libraries rely solely on state funding. A reduction of this magnitude could mean many will close branches or reduce hours and services.”

Mentor Library’s Board of Trustees responded to Governor Strickland’s proposal with a resolution directing the library administration to seek and implement technology to make up for the budget shortfall (Mentor Public Library Board of Trustees, 2009, p. 2). To this end, self check-out stations were installed in the Main Library during the summer of 2009. As he watched the installation of these stations, and faced with the potential of unstaffed reference desk time, it occurred to David Newyear (at that time Manager of Adult Information Services at MPL) that there might be a way to create self-serve information stations as well.

Having seen successful implementations of AI Chatbots in corporate and public sectors, David did some research and began creating a chatbot using SitePal (a company which provides animated avatars) in June of 2009. Emma, the Mentor Public Library’s virtual agent, made her debut on the library’s website on November 19, 2009. This first iteration of Emma provided answers to a list of twelve “frequently asked questions” or about the library’s services and policies. In January 2010, Sitepal’s artificial intelligence component was enabled allowing Emma’s programming to begin to evolve to address more and more complex questions. During the month of February, Emma answered nearly 10,000 questions posed by patrons.

From March to May of 2010, Emma’s “brain” was migrated from SitePal to the Pandorabots website, and Emma was rebuilt to take advantage of Pandorabot’s Superbot 2.0 base AIML files. At this point Emma was answering approximately 300 questions each week with a correct response average of 60%. In July 2010, David invited Michele McNeal, the Web Specialist for the Akron-Summit County Public Library, to collaborate on the project and they began to work together to expand the scope of the bot’s capabilities and improve the correct answer ratio.

Complete Chapter List

Search this Book: