Content area
Full Text
Celebration issue
Introduction
Technology is changing the way we communicate. Today's students thrive on social networking tools like Twitter, Facebook and Foursquare. Chatbots are not usually included in this grouping but they engage users with a playful interface that is familiar to a generation that grew up with online games. On a recent posting by Beloit College about the incoming class of freshmen who will graduate in 2015 (www.beloit.edu/mindset/2015/): "They've always had the privilege of talking with a chatterbot." Libraries are seeking ways to engage this generation and should consider the chatbot as another tool for reaching users who expect more than a flat website.
This paper is about adapting artificial intelligence technology for reference services. AI has come a long way as IBM's Watson demonstrated when it won "Jeopardy" in February of 2011. The advances in artificial intelligence (AI) combined with the availability of online resources make it time to consider artificial intelligence as a tool for the library.
Chatbots (also known as conversational agents, artificial conversation entities, or chatterboxes) are computer applications that imitate human personality. A chatbot is interactive, responding in sentences that track the conversation in a way meaningful to humans. This characteristic of mimicking discourse appeals to library users who want a more interactive library experience, something livelier than a search engine, and fits well with the socially directed students we are seeing on our campuses.
One of the selling points for these bots is their ability to handle common directional and predictable questions. They excel at routine, repetitive tasks that can free librarians from the most common questions. Bots flatten a website; when someone chats with a bot, they do not need to know the layout of the website, or the resources available to them. The chatbot is programmed with that information and pulls together the necessary sources, reformatting and presenting it in a manner that meets the needs of the information seeker.
Can a chatbot truly replace the experience someone gets in a reference interview? Is it even possible to identify the best characteristics of a reference experience and develop a program algorithm that will reproduce that experience? In a 1996 study undertaken by [1] Nardi and O'Day (1996), the authors analyzed the activities of face-to-face reference sessions...