February 2019
Social Media Analysis – What Facebook Tells Us About Social Cohesion in Sri Lanka
The briefing paper produced by DRI in collaboration with Hashtag Generation provides an analysis of how social media influences Sri Lanka’s political discourse.
Graphics on Monitoring Social Media in Elections
More and more organisations and researchers start monitoring social media in elections. The graphics prepared by Rafael Goldzweig with input from Michael Meyer-Resende provide orientation for such work.
August 2018
An opinion article on POLITICO by DRI´s Raymond Serrato and Michael Meyer-Resende on how Facebooks Al-based approach fails in conflict countries.
Is Facebook a destabilising force? – BBC Newsnight
A recent investigation of Facebook by Reuters revealed more than 1,000 anti-Rohingya posts calling for their murder, months after Facebook claimed to have improved its moderation of the site. Newsnight, BBC’s flagship news and current affairs TV programme, interviewed DRI’s Senior Programme Officer, Raymond Serrato, to learn more about his research on hate speech and social media in Myanmar and whether Facebook has done enough to address the problem.
April 2018
Revealed: Facebook hate speech exploded in Myanmar during Rohingya crisis
DRI´s Senior Programme Officer and Governance & Innovation Expert Raymond Serrato “examined about 15,000 Facebook posts from supporters of the hardline nationalist Ma Ba Tha group. The earliest posts dated from June 2016 and spiked on 24 and 25 August 2017, when ARSA Rohingya militants attacked government forces, prompting the security forces to launch the “clearance operation” that sent hundreds of thousands of Rohingya pouring over the border.” Guardian
March 2018
Fighting fake news: caught between a rock and a hard place
The discussion on disinformation (“fake news”) turns increasingly towards the question of how to prevent the phenomenon in the first place. The author of this opinion piece argues that prohibitive measures run the risk of either being ineffective (if very restrictive in their scope) or of being subjective and politically motivated (if too broad in their scope). He calls for distinguishing between technical cyber warfare threats (which should be prohibited and addressed directly) and the cognitive approach mis-using social media’s tendency to reward polarizing and extreme content to create disinformation campaigns (which should be addressed indirectly via digital education and transparency measures).
February 2018
In this piece, Nina Jankowicz looks at the potential solutions to disinformation campaigns and other negative effects of social media for democracy. She surveys existing innovative initiatives to teach digital and media literacy to citizens from around the world and argues against too much trust in what she considers quick fixes, such as de-bunking. Instead she advocates for investing in people’s abilities to deal and discern information received on the internet.
Why the hysteria around the ‘fake news epidemic’ is a distraction
In this opinion piece, the political scientist Cas Mudde makes the case for focusing on improving quality reporting of traditional media rather than getting distracted by the fake news phenomenon, which, as recent research shows, is more limited in its reach and thereby also in its impact than hitherto feared. He also warns that the policy measures discussed to counter fake news could – in certain, authoritarian contexts, be used against the freedom of the media as such.
Harvesting Social Signals to Inform Peace Processes Implementation and Monitoring
This study uses Twitter data ahead of the referendum on the Colombian peace process in 2016 to predict the outcome of the referendum and to show that public sentiment on Twitter indicated a negative outcome for the referendum (different from public opinion polls). The study highlights the use of social media analytics for early warning systems on social conflict as well as as alternative to opinion polls for predicting election outcomes.
This report investigates how important “fake news media” is by looking at a number of online media in France and Italy. In both countries those websites which regularly carry fake news articles are much less widely read than the most prominent mainstream online news media. However, a few fake news sites are shown to create a disproportionately high number of interactions on their site, suggesting that, in some cases (i.e. for some news sites and some of their stories), fake news can indeed go viral. The article also discusses important methodological limitations of the chosen approach.
The Guardian. “‘Fiction is outperforming reality’: how YouTube’s algorithm distorts truth.”
This article sheds light on the role that YouTube may play in forming political opinion. While a lot is being written about the role of Facebook and Twitter on democracy, YouTube has escaped scrutiny. The article argues that YouTube tends to recommend extremist or sensationalist content to keep viewers glued to the service, with negative effects on democracy. It cites research that suggests that in the US Presidential elections YouTube was far more likely to recommend pro-Trump film clips than those that would favour Clinton, because the Trump clips better served the algorithmic interest in sensational content.