An Examination of Public Discourse on Human Gene Editing Using Natural Language Processing

This research aims to explore the different ways in which scientists, ethicists, journalists, and commissions speak to the public about new gene-editing technologies. The research collected more than 100,000 sentences from books, news articles, and reports written by these four types of authors, examined the relative distinctiveness of their speech, and compared the prevalence of key terms among the four groups. In addition, the sentiment of each group's speech was compared using IBM Watson's sentiment classification functionality. Key findings suggest that some topics-such as the issue of "enhancement," the parent-child relationship, or the role of China-are covered very disproportionately by different groups of authors, and that the current discourse regarding gene editing has not demonstrated extensive consideration of the perspectives of the disabled, religious adherents, or other relevant interest groups. While some group-level differences are highly robust, on many measures, the variation within groups is wider than the variation between groups.