From the ‘I am an immigrant’ poster campaign, a Joint Council for the Welfare of Immigrants and Movement Against Xenophobia effort to “challenge the negative rhetoric against immigrants, celebrate them and provide their story.” (Broad sharing policy) Recently, Facebook's confirmed sharing of at least 87 million profiles with Cambridge Analytica has made big headlines, especially in connection to the US. But reports of this collusion have been in the news for some time in the UK, particularly in relation to the 2016 national referendum to leave the European Union otherwise known as “Brexit.” In journalist Carole Cadwalladr's words last year, democracy itself was “hijacked” through Cambridge Analytica operations; her report called it the “Great British Brexit Robbery” (a report that is still the subject of legal complaints). To what extent technological platforms have been used to shape public response is a matter of strong concern, as with the case of Brexit. Anyone who has been paying attention to Britain's departure from the European Union next year knows that it is a topic that provokes intense emotion. So, is there something in the topic of Brexit that disinformation efforts can take advantage of? What perspectives might a high-level, data-driven analysis of Brexit news provide? According to a Media Cloud collection of US and UK media mainstream sources, and a web crawl of outlets connected to them, the topic of Brexit has appeared in approximately 70,000 stories between 1 March 2017 and 28 February 2018. What is Media Cloud? Media Cloud is an open source platform developed by the MIT Center for Civic Media and the Harvard Berkman Klein Center for Internet and Society. Media Cloud is designed to aggregate, analyze, deliver and visualize information while answering complex quantitative and qualitative questions about the content of online media. Major themes that emerged in Media Cloud included a number of expected key terms such as references to the US election and the UK referendum and prominent politicians such as Teresa May or Boris Johnson. However, when focusing on topics related to potential motivations behind Brexit among the terms, the issue of the economy appears to be more frequent. In addition, a prominent topic of discussion in the media was immigration. This is an ordered word cloud of a sample of stories that contained “Brexit” in Media Cloud. The words that show up more often appear bigger and show up first in the list. (View larger image) Exactly how were these topics discussed in US and UK media in relationship to Brexit? To get a sense of the tenor of the conversation around immigration, our Bias Prism tool processed key terms related to the discussions around immigration. Bias Prism is a Natural Language Processing tool that analyzes language for expressions of personal perspective and potential bias. What is Bias Prism? Bias Prism is an experimental feature being developed by the Georgia Tech's Behavioral Modeling and Computational Social Systems Group in partnership with NewsFrames. Results from Natural Language Processing algorithms signal a number of possible ways that texts may be using perspectival or biased language. Learn More >> The Bias Prism tool aims to offer more precise ways of thinking about the presence of perspective or bias in statements. Rather than producing a “biased/not biased” result, researchers are able to analyze texts through a number of considerations such as sentiment or expressions of doubt. Comparing the samples of key terms flagged for further inspection, stories talking about immigration used more terms — like “damaging,” “anxieties,” “crisis” — that signaled perspectives and emotions than were used when discussing the economy, even though there were many more sentences about the economy. Exploring the stories in context further, the key terms of immigrant or immigrants also came to the fore. They followed the same pattern, with Bias Prism results around immigrant appearing slightly still more perspectival or biased than those around immigration. Our examination of several Media Cloud samples examined 4,500 words. Based upon a statistical analysis of Bias Prism results, keywords related to “immigration” and “immigrant” showed significant tendencies towards more perspectival or biased language. The averages for each sampling — Brexit keywords overall, focused around the economy, around immigration, and around immigrant — were progressively higher. Higher numbers indicate the potential for more perspective or bias.