Cyborg Politics: A Study of Artificial Agents in Online Democratic Deliberation
Artificial agents in the shape of social bots are an increasingly important feature of communication on the internet. There is increased use by political campaigns, militaries, government-contracted firms, corporate lobbyists, and civic activists, of social bots in online attempts to disrupt political debate and to manipulate public opinion.
Politicized social bots — i.e. political bots — are used, for example, to flood social media streams with spam during political crises, conflicts, and elections in order to interrupt political efforts of other actors, or to boost follower numbers and response levels to produce false impressions of popularity. They can also be used to send out sophisticated computational propaganda, and to feign grassroots movements.
The common everyday occurrence and potentially huge consequences of political bots makes analyzing and understanding their role and functioning a priority for social science. Previous scholarship on various forms of social bots has either tended to be largely theoretical, or one-sidedly technical focusing for example on bot creation, bot prevalence, or bot detection. With this project, we argue that any attempt to assess the impacts of bots on society and politics must start from an empirical, social, and contextualized perspective.
We use a mixed-methods approach to carry out a sociotechnical analysis of political bots. The goal is to produce knowledge that can help better address and understand the political role and consequences for democracy of these artificial agents. The analysis will be carried out at four analytical levels:
• Communication: What do political bots say?
• Connections: In what social network relations are political bots involved?
• Context: How does society respond to the presence of political bots?
• Consequences: What are sustainable ways for society to deal with challenges for democracy that are raised by the increased use of political bots?
Professor of Sociology, DIGSUM, Umeå University