Lockerd, A. "Understanding Implicit Social Context in Electronic Communication"

Abstract

Artificial Intelligence (AI) has shown competence in helping people with complex cognitive decisions like air traffic control and playing chess. The goal of this work is to demonstrate that AI can help people with social decisions. In this work Artificial Intelligence of Social Networks is used to improve human-human communication, recognizing the social characteristics of human relations in order to achieve a more natural online communication interface. Can a computer learn to understand the value of communication? It is shown here that a first attempt at social context classification performs with almost 70% reliability. Could a computer use this to help a person relate to other people through technology? The addition of social context to an email interface is shown to have a positive effect in a user's online communication behavior.

Email is a tool that people use practically every day, making an implicit statement about their relationships with other people, and providing an opportunity for a computer to learn about their social network. Furthermore, over the years people have come to utilize and depend on email more in their daily lives, but the tool has hardly changed to help people deal with the overwhelming amount of information. Many of the social cues that allow people to naturally function with their social network are not inherent or obvious in Computer Mediated Communication (CMC). This work offers automatic social network analysis as a means to bring these cues to CMC and to foster the user's coherent understanding of the people and resources of their communication network.