We are right to be worried about the polarization of public debate, the rise of populism and digital propaganda. It also goes without saying that social media have a growing impact on our politics and society. However, one should be cautious not to confuse observations with explanations. Focusing on fake news, bots and algorithms it is easy to miss real agents behind the screens: humans. Not only those who create content and tech tools, but also average users. Can we, as individuals, control and influence the quality or diversity of information we receive online? Is it our responsibility to “consume responsibly”? Are we forced to live inside information bubbles or can we do something about it?
Seeking answers to those questions, Panoptykon Foundation invited researchers from the University of Pennsylvania – Dr Emad Khazraee and Pawel Popiel – to design a case study of political debate on Polish Twitter (social media venue most populated by political influencers and journalists).
We wondered who was creating trends, who spoke to whom and what type of conversation it was. Did influencers with different opinions confront each other? Was the debate manipulated or otherwise influenced by false amplifiers? In September and October 2017, when Polish streets and social media venues were fuming with civic unrest (women protesting against ban on abortion, young doctors fighting for public healthcare reforms, citizens defending independent judiciary), we collected and analysed nearly one million tweets.
Here is just a snippet of what we found out:
- Polish Twitter does not encourage confrontation of opinions. In both political bubbles that we identified influencers talked about the same topics but hardly ever talked to each other.
- It is prominent individuals and well-established organisations (politicians, journalists, mainstream media) that shape political discourse on Twitter, not bots. Bots (false amplifiers) can be used for influence-boosting but on their own they are not in a position to change or create trends.
- Using network analysis (without deeper, qualitative analysis) it is nearly impossible to differentiate false amplifiers from professionally managed accounts.