Applied Text Analysis with Python

Chapter 10. Chatbots

In this chapter we will explore one of the fastest-growing language-aware applications: conversational agents. From Slackbot to Alexa to BMW’s Dragon Drive, conversational agents are quickly becoming an indispensable part of our everyday experiences, integrated into an ever broader range of contexts. They enhance our lives with extended memory (e.g., looking stuff up on the internet), increased computation (e.g., making conversions or navigating our commute), and more fluid communication and control (e.g., sending messages, managing smart homes).

The primary novelty of such agents is not the information or assistance they provide (as that has been available in web and mobile applications using point-and-click interfaces for a long time); rather, it is their interface that makes them so compelling. Natural language interactions provide a low-friction, low-barrier, and highly intuitive method of accessing computational resources. Because of this, chatbots represent an important step forward in user experience, such as inlining commands naturally with text-based applications thereby minimizing poorly designed menu-based interfaces. Importantly, they also allow new human–computer interactions in new computational contexts, such as in devices not well suited to a screen like in-car navigation.

So why is this rise happening now, given the long history of conversational agents in reality (with early models like Eliza and PARRY) and in fiction (“Computer” from Star ...