With documentaries like ‘The Great Hack’ we are deeply fearful – and rightfully so – of how our data can be used, abused and weaponised. Over the past three years, significant political developments on both sides of the Atlantic have fuelled a global debate around the role, reach and relevance of social media in our daily lives including every imaginable aspect of our economic, political, social, media and even cultural interactions. While this anxiety is generally grounded in mature studies of cause and effects, many extrapolate lessons for Sri Lanka that aren’t entirely valid and even downright dangerous. While documentaries on Netflix serve to highlight critical issues, they are terrible at guiding policymaking that responds to a class of problems academics called ‘wicked’. A ‘wicked problem’ is one that has no easy resolution and where the very act of intervening or even observing changes the nature of the problem, very often to an even greater level of complexity. So in sum, while there are definitely problems around the use and abuse of social media as it is popularly referred to, could there also be benefits to hold those in power accountable, despite their best efforts at resisting this, and interrogate, for example, the behaviour of candidates and political parties in an election campaign? If in a country like Sri Lanka which doesn’t have legal frameworks in place for the sunlight of accountability or scrutiny around campaign finance and spending, can dynamics and data determined on social media, once harvested and analysed, give clear pointers around how online propaganda have impacted voter behaviour and election outcomes? And in Sri Lanka, what constitutes ‘social media’? And where is politics online? Does it lie in domains openly affiliated to or controlled by politicians and political parties? Or has it migrated to gossip, meme and other online spaces that are more generally defined through religious or nationalist fervour, with no overt connection to a partisan agenda or candidate? And when we talk of social media, do we embrace new apps like TikTok, or instant messaging? If we don’t know, and know not how to even frame key questions, how can we expect policymakers to protect citizens from a degree of electoral manipulation they may not even realise exists?
This is the landscape I live in, travel around and study, every day. And over the next weeks, as I have in the past, the attempt through this column will be to share more widely data that can inspire questions and conversations that are anchored to data and evidence. Arguably, intellectual debates on the thrust and timbre of politics is a luxury in any electoral process, where spectacle and emotion hold sway. In Sri Lanka, the chest-thumping populism of leading candidates will organically result in digital echoes in addition to engineered popularity – or in other words, the appearance of mass appeal. This isn’t new and is textbook propaganda for decades clearly evident in our newspapers, TV and radio. But here’s the thing. Social media allows for new possibilities around the manufacture of propaganda, for example, on Instagram – a photo-sharing app. Don’t believe me? A study this week of the Instagram accounts of the Prime Minister, UNP, two accounts of Sajith Premadasa (one launched specifically around his candidacy bid for the presidential election), Mahinda, Namal and Gotabaya Rajapaksa reveals tens of thousands of followers and hundreds of thousands of engagements in the form of comments, pushback, love and endorsement. The personal is political. Significant differences between these accounts, never before studied in Sri Lanka, clearly frame very different political projects. Captions of the photos range from what government and party have done to eco-centric framing, anchored to self and personality. Some highlight party and fellow MPs more than personal endeavours. Others do not. Then there is the question of volume. Over past 30 days, Namal Rajapaksa’s Instagram account got 3 times more engagement than Gotabaya Rajapaksa’s account, 6 times more than his father Mahinda Rajapaksa’s account, 43 times more than Sajith Premadasa’s presidential campaign account & an astonishing 88 times more than incumbent PM, Ranil Wickremesinghe. Instagram is very popular amongst a demographic too young to vote, but will in a few years. It stands to reason that those who are on the app will also vote for the first time in the presidential election. How they see politics through this one app will inform how appealing they find candidates and the electoral process in general.
The manipulation of preference towards individuals is one thing, but another danger is around the manufacture of apathy around franchise. In the past, voter suppression happened by preventing voters from physically accessing ballot boxes. Today, the same effect for a younger demographic can be attempted by the manipulation of emotions over the apps and platforms they engage with, in ways that are cloaked.
And what of traditional media embedded in social media? During the constitutional crisis, traditional TV channels had, by far, the most visible footprint on Facebook. Their videos were viewed tens of millions of times. Social media in Sri Lanka today isn’t the domain of liberal cosmopolitan civil society in and around Colombo, as it is popularly projected and resisted. If anything, the data confirms the contrary – civil society’s footprint on social media is negligible. In other words, the promise of social media to provide alternative and critical perspectives to mainstream or traditional media is risible in the main. In Sri Lanka, the same dominant narratives in the media my generation grew up with has a vice grip on the framing of issues over social media. For example, I study over 1,600 pages on Facebook anchored to politics, religion, popular culture and media. Data around how audiences engaged with the SLPP convention, Anura Kumara Dissanayake’s event at Galle Face, and Sajith Premadasa’s rally in Galle – again, not something anyone’s done before – highlighted the fact that one of the candidates benefitted enormously from the coverage given to his rally by a prominent TV channel, which live-streamed it on Facebook. I published these figures on Twitter because it is essential for voters to know what channels support which candidates. The JVP for example, or any independent candidate, will not enjoy this level of support by partisan mainstream media, placing them at a distinct disadvantage, even with the affordances of social media to reach voters independently.
Also last week, I studied how on YouTube, the algorithm itself recommended videos based on a search pegged to Sajith Premadasa, Gotabaya Rajapaksa or Anura Kumara Dissanayake. The endeavour here was to determine to what degree algorithmic bias impacted the framing of a candidate. Turns out that two TV channels to varying degrees dominate the framing of these three individuals. Since the YouTube algorithm basically reflects the volume of and engagement with content uploaded to the site, this, for the first time, gave an insight into how polarisation can be exacerbated as a consequence the nature of content around an individual or issue, which algorithms go on to amplify. Is it social media regulation that is needed? Or is the carcinogenic rot in mainstream media, well-known and studied, to blame? Is this a problem of technology, or a partisan bias and dependence on political favours that predates Facebook? Answers to these questions are there in the data.
All the observations above were placed in the public domain, in great detail. Though I have my own bias and preferences, it matters less to me who wins the presidential election that how it is conducted. Politics in the form of newspaper ads, radio or TV spots, posters, rallies and the gate to gate, face to face variety will endure. But there is an industry that is both worth millions, and engages tens of millions, which exists on social media. Not all of it is by design dangerous. In studying these new influence engines, curated by professional agencies, it isn’t always clear who is responsible for what. But like potholes on a road that impede traffic flow or result in detours, data allows us to determine how ideas are projected, anxiety is augmented, fear is fomented, anger is amplified, ideology is highlighted, populism is projected, history is erased and desired futures, carefully crafted. It is in equal parts fascinating and frightening. This is stuff that shouldn’t really be limited to my Twitter feed or this column. It is integral to how we must engage with and resist propaganda.
Nothing less than the quality and nature of our democracy is at stake.
###
First published on 1 September 2019 in The Sunday Island.
Sanjana Hattotuwa 