Re ection Machines Requirements to prevent the e ects of micro-targeting in political scenarios
Over the past decade, the focus of political campaigns has shifted more and more towards social media and the use of technology. With the use of psychographics, micro-targeting can be used to manipulate voters towards a decision. One of the most recent examples for this application is the Cambridge Analytica (CA) scandal in the 2016 US elections. They sent out personalized messages to indecisive voters to manipulate them, which seemingly helped Donald Trump win this election. This thesis aims to provide a speci cation of a re ection machine (RM) that is able to prevent those manipulation e ects by presenting re ection messages to a user. RMs are similar to decision support systems, however, they do not aim to provide a correct solution to the user but enhance the re- ection process by proposing a question or statement questioning the user's decision. In the domain of political voting scenarios the re ection process is not enhanced by questioning a decision but by raising awareness of possible micro-targeting proactively. It is examined which requirements such RM could have when it is to be implemented. This entails the types of questions, the information needed and the role of the user's information. To demonstrate the research, a thought experiment with conceptual users from di erent personality spectra is presented and a mock-up simulation of an RM to show the enhancement of the re ection process is executed. The evaluation shows that there is potential in RMs in relation to raising awareness of micro-targeted political messages, however more research is yet to be done in this domain.
Faculteit der Sociale Wetenschappen