Tracking the Peace: Analysis of “Peace Speech” Using Machine Learning and Artificial Intelligence

Abstract: The properties and content of speech can represent and reinforce the social processes that generate and sustain “positive peace” in a culture. New technologies are now available to analyze text, speech, and video to better understand the characteristics of such “peace speech”. We describe how these new methodologies of machine learning and artificial intelligence can be used in social science. We show how we have used those approaches to identify and study the linguistic properties in on-line news media that differentiate lower and higher peace countries. From that work we also developed computational tools to analyze on-line news data to measure the level of peace in a country. We are now using artificial intelligence to identify the social processes that underly the peace speech in peaceful countries. These studies may also lead to real-time displays of measures of peace and predictive analytics of future changes in the levels of peace.

Keywords: sustaining peace, Artificial Intelligence, Machine Learning, journalism