Seeds to blossoms: Social media post-war (in Sri Lanka)

I was invited with colleagues from the National Centre for Peace and Conflict Studies (NCPACS) to deliver a presentation on our research and motivations to a group of locals this week, assembled at the Mornington Methodist Church in Dunedin.

The central challenge around presenting my doctoral research for many who are fearful or anxious about social media in the main, don’t know the Sri Lankan context, particularly post-war, and aren’t remotely familiar with data science, is on how to best communicate, in a way that sparks interest and curiosity, what I do and why I do it.

As an early slide of mine showcased, half-jokingly, this endeavour is ironically more difficult in a community, context and country that is the second most peaceful in the world, and where local media’s reportage of violent conflict is around ducks being shot in the back (yes, seriously).

I have used fire, water and nature as a way to help explain the genesis, growth, spread and engagement of content over social media. For this presentation, I used the bloom and blossoming of cacti and a rose.

Going through snapshots of what Facebook usage is like in both Sri Lanka and New Zealand (biggest difference really is that in New Zealand, iOS users are much more than in Sri Lanka) I noted that I specifically will not be showing any photos of violence, war and trauma from Sri Lanka because not only are they a Google search away, my endeavour was to frame solutions, not sell the worst of what we are and have been.

A brief introduction to social media frames my research was partial to prefaced, for me, the key slide of the entire deck, which had looped videos of nine flowers in bloom.

Most of the flowers (Echinopsis Cacti to be precise) were taken from this Vimeo video, to illustrate how on and over social media, content and conversations evolved.  I likened conversations online – their genesis, spread, key proponents, timbre, form, reach and contours – to the bloom of flowers. Each flower (or conversation) distinct, yet in key ways, similar or connected. I also noted that with cacti and roses, what looks immediately appealing can hide thorns and sharp shards, unseen at first, but present and alive within, under or as part of the blossoming. Some conversations, like flowers, I said stayed alive more than others. The life of other conversations, I said, also like the environment or flower bed a plant grew in, was entirely dependent on context, climate and culture.

I followed this slide with my own data visualisation of the Jana Balaya Rally and #stillnoanswers, Amnesty International’s campaign on the disappeared, both of which happened around the same time in Sri Lanka. Without going into too much detail, I noted how research into conversational dynamics over social media, accounting for identity frames, language, geography, platform affordances, device limitations, network effects, education and economics, was not too different from how a gardener would tend, lovingly and with great care, to a flower bed.

Just a glance, I noted, was enough to reveal that one movement or campaign was very different to the other. As a researcher, I noted, my motivation was to ascertain why.

Screenshot 2018-10-19 at 8.27.33 AM.jpg

I then showed snapshots of Twitter based on my preliminary observations and data collection. I flagged how around the violence in Digana in March, there was a discernible increase or spike in the content generation on Twitter alone – leading up to the kinetic violence on the ground, and during it. I flagged that causal linkage between the volume of content and a violent context, if any, was undetermined as yet.

Screenshot 2018-10-19 at 8.26.49 AM.jpg

I then gave two snapshots of Facebook, again from data pulled from my research from just 1st January 2018. Noting that the dominant drivers and primary frames on politics, country, context and the world were gossip sites in Sinhala, I noted that the media diet of a young demographic, post-war, was akin to always or only eating fast-food. I then flagged how much greater engagement the extreme nationalist Sinhala-Buddhist pages and accounts I monitored gathered over accounts pegged to civil society.

Screenshot 2018-10-19 at 8.26.44 AM.jpg

I ended up with frames I would employ in my doctoral research around social media analysis. Using a video of bubbles, I noted how many seemingly disparate conversations online were, based on language, locale, issue or proponent, connected, with these connections morphing over time. Using the awful hate directed against Sandya Eknaliyagoda as an example, I noted that the conversational domains I often looked at (and am sometimes an actant in or inadvertently thrust into) were very far removed from decency and civility. I noted that my interest was in ascertaining the drivers of hate and violence, that after finding expression online, could exacerbate, render more intractable, lead or contribute to violence in the real world.

I ended with a quote from Voltaire, discovered by reading (the absolutely wonderful) War Gardens: A Journey Through Conflict in Search of Calm by Lalage Snow. The quote I said captured the primary motivation for my doctoral research – to look at home and my work since 2002, as fertile terrain to help unpack why and how conversations online are the way they are today, and how, if at all, their thrust, timbre and topics can be changed.

Screenshot 2018-10-19 at 8.44.35 AM.jpg

My hope, I said, was that one could eventually extrapolate from my work on and in the post-war context in Sri Lanka ways of looking at and dealing with hate online in fragile democracies, applicable or resonant elsewhere.

Download the full presentation as a PPT here.

Trolls, bots and digital propaganda: What it all means, and how it will impact citizenship

This blog post should have come up much earlier than this, but Digana got in the way and in a way, resonating with what was discussed, esp. through a weaponisation of Twitter hitherto unseen in the kind of violence that gripped Sri Lanka in general, and Kandy in particular, that week.

After Namal Rajapaksa, bots and trolls: New contours of digital propaganda and online discourse in Sri Lanka, first published on Groundviews went viral, I was inundated with emails seeking more information on what we had written on and warned against. The requests for clarification, more information on and ways to safeguard against what we had noted came from civil society, and perhaps unsurprisingly, also from sections of government.

Instead of responding to each and every one, I decided to have an open forum to discuss the article and key issues arising from it.

Invited Yudhanjaya and Sabrina Esufally, from Verite Research, to join a panel that was moderated by the Executive Director of the Centre for Policy Alternatives (CPA). Yudha and Sabrina gave excellent presentations, looking respectively at the technical aspects (the data told a compelling story) as well as the long-term impact and potential solutions to the challenges we had outlined.

All of our presentations and related material are online here.

My presentation captured in brief an almost total reversal in the perception of social media as democratic, emancipatory platforms that helped citizens overthrow illiberal, dictatorial and authoritarian regimes to over the past year alone, its unprecedented weaponisation by non-state and government actors. This is in addition to disturbing practices by the social media companies themselves. I flagged individuals like Brad Parscale in the US and Davide Casaleggio in Italy as individuals, with supremely capable minds adept at political communication over social media, now as powerful as, if not more so, than the politicians and political parties they worked for or with.

Flagging the weaponisation of Facebook and how countries like Sweden were taking measure to safeguard critical infrastructure elections from foreign interference (read Russian psy-ops and hacking), I gave an overview of what Yudhanjaya and I discovered around Namal Rajapaksa’s Twitter account. I went into some detail to explain what a troll and bot were and how they polluted discourse online, as well as how cheap they were to deploy.

Providing frames of entry for both Yudha and Sabrina who spoke after me, I flagged some of the topline data and a meeting with Sri Lanka’s Elections Commissioner, Mahinda Deshapriya, late last year, around some of the issues other countries in the West were preparing for, safeguarding against, dealing with and as I submitted, Sri Lanka also needed to take very seriously.

I warned against a troll or bot arms race, where others seeing what Namal Rajapaksa had done and how, also would try to do the same and better. This I said would lead to a situation, backed by Yudhanjaya in his presentation, where public discourse would be completely overrun by narratives, ideas, frames and perceptions determined by parties who were adept at manipulation, subterfuge and deception. I flagged regulation as way forward, but with many pitfalls too, if championed and overseen by a government with poor democratic principles.

The threat to the timbre of public discourse and democratic institutions is real, as the growing partisanship in the US clearly shows. Given the violence in Ampara that had just occurred (our conversation pre-dated the catastrophic violence in Digana by just a day or two) I noted how an attention economy can and will be gamed, clearly brought out by the viral appeal of a video first shared on Facebook around the sterilisation pill myth that fuelled the violence.

I also flagged how Namal Rajapaksa deleted inconvenient tweets from the past that painted him in a light incompatible with the reinvention of self and branding he was undertaking at present. This I submitted was more generally applicable to social media, and strategies employed to whip up unrest, and then delete all traces of the material that fuelled the violence, or was the cause of it.

I quoted George Marshall’s speech from 1947, launching the Marshall Plan, twice, because of how resonant it is in our contemporary information landscape, and the challenges arising from it. My penultimate slide called for a much greater emphasis to be placed on media and information literacy, from the time children entered school. Sabrina’s presentation built on this considerably and is well worth taking a look. The entire panel was in agreement that critical media appreciation was vital as a means through which to really address the problems around rumour and misinformation taking seed and growing over social media. But we also agreed more urgent measures needed to be taken in order to address the immediate challenges posed by the weaponisation of social media in Sri Lanka.

And then, Digana happened.

Namal Rajapaksa, bots and trolls: New contours of digital propaganda and online discourse in Sri Lanka

In the last quarter of 2017, pushback over Twitter to content Groundviews pushed out over the same platform came from sources not encountered or interacted with before. This piqued the interest of the site’s founding editor, Sanjana Hattotuwa, for one key reason. All the accounts publishing content against Groundviews were overwhelmingly promoting and partial to Namal Rajapaksa, a Member of Parliament and the extremely (social) media savvy son of the former President Mahinda Rajapaksa.

The troll army retweeting and promoting Namal Rajapaksa’s Twitter account was overwhelmingly anchored to profile photos that were fake, and registered to names that deviously sounded like they were from the Muslim, Tamil and Sinhala communities, but were also fake.

In any case, the data clearly suggests Namal Rajapaksa drawing a highly predictable number of followers on to Twitter every day.

What’s interesting for social media research is the manner in which the @RajapaksaNamal account on Twitter is used, or arguably, abused. It reflects a new appetite for social media strategies specifically engineered for electoral gain amongst all politicians, and not just the Rajapaksas and Joint Opposition, involving human trolls as well as automated bots. The intent it clear – to influence voter perceptions and public discourse, over and beyond social media.

…the danger around the weaponisation of social media around electoral processes is that neither government nor civil society is prepared to deal with it.

…what is now a danger is that the followers (in the form of bots and trolls) can also be strategically leveraged to quell dissent, shape narratives, highlight propaganda, spread misinformation, drown out critical voices, bully, act as echo chambers and shape social media discourse.

Without sounding alarmist, Sri Lanka has already entered a new online political dynamic, in which the discursive landscape is governed agents of censorship, manipulation and control outside the parameters of traditional observation and analysis. This isn’t just a technocratic concern.

Co-authored with the amazing Yudhanjaya Wijeratne, read the article in full on Groundviews here.

Social media, viral news and the future of peace negotiations: Panel at Build Peace 2017

I was invited to take part, over Skype video, in a panel on social media, viral news and the future of peace negotiations at Build Peace 2017, held early December in Bogota, Colombia. My great disappointment at not being able to attend in person was somewhat offset by what was a great conversation with Juanita León, Director, La Silla Vacía moderated by the inspiring Helena Puig Laurri, co-founder and co-director of Build Up.

I don’t think there was a video recording of our session, which was in itself no mean technical achievement, since I was connecting was Sri Lanka, Juanita from elsewhere in Bogota and Helena at the conference venue, patching us all in. Skype, when it works, is quite incredible.

In brief introductory comments, I flagged my work in 2002 with the Sri Lankan Ceasefire Process and the One-Text process for which I designed technical solutions for asynchronous, confidential, encrypted Track 1 and Track 2.5 negotiations and communications using, what was at the time, commercially available off the shelf software (called Groove).

A lot of this work, and subsequent research, pre-dated and prefaced, by many years, the current praxis and increasing academic research on the use of technology for peacebuilding.

Referencing my keynote address at the first Build Peace conference three years ago, I spoke about radical transparency and how combined with now ubiquitous computing devices that recorded passively, or actively through user intervention, their surroundings, what it meant for negotiations processes where an element of timing, founded on secrecy, remains essential. I flagged how we needed to re-evaluate the meaning and effectiveness of the Chatham House Rule, for example, and how to engineer today a process that retains confidentiality in a manner that doesn’t also infringe on the rights of participants in a process to speak aloud and freely about what they are doing, and involved in.

I noted that with radical transparency also came radical inclusion – the idea that everyone today was a stakeholders in a negotiations process, with what happens in society and polity deeply and almost immediately impacting high-level discussions, as well as vice-versa. My challenge was around how to create processes that included stakeholders who had the power to bring about change and exclude spoilers and spoiler dynamics, which isn’t as easy as it was even just a few years ago – and certainly at the time I engineered the platforms for the One-Text process in Sri Lanka. The management of expectations and optics has I argued a substantive bearing on the negotiations.

I spoke about how difficult it is, ironically, to ascertain the interests of negotiating partners because despite the tsunami of content created, it was increasingly difficult to tune out the noise and focus on the signal. Given filter bubbles inhabited by those at the table, the challenges around misinformation, disinformation and sophisticated, web based propaganda campaigns, it is no longer possible to understand and respond to stakeholder positions, even as the manipulated of public opinion is easily possible with social media.

I flagged the importance of information and media literacy in this regard, noting that for example in Sri Lanka, while adult literacy was very high, people actually believed what they consumed over social media, leading to a disturbing situation where rumour and misinformation online and over instant messaging apps stood to derail fragile negotiations processes more than help secure them. I also flagged how power, patriarchy, social, cultural and political norms were ingrained into what is increasingly algorithmic bias – meaning that the social media platforms we use daily to help understand the world outside are themselves hostage to algorithmic filtering that only promotes what one is already partial to – increasing over time, and not bridging, differences between communities and regions.

In light of what I’ve noted elsewhere publicly, I also wondered how psychometric profiling, sentiment manufacture, trolls and the use of bots, the public mood could today be influenced in subtle (or not so subtle) ways that had a direct and lasting bearing on complex negotiations.

I wondered how the constant chatter – often without the art of the long view, vision, reflection, context or calm contemplation – impacted the reception and perception of what was publicly known of negotiations, putting those in the process under intense pressure to not just negotiate across the table, but also almost in real time, with their own constituencies as well.

I quoted Dannah Boyd, a well-known researcher from Microsoft ,

“[W]e have a cultural problem, one that is shaped by disconnects in values, relationships, and social fabric. Our media, our tools, and our politics are being leveraged to help breed polarization by countless actors who can leverage these systems for personal, economic, and ideological gain.”

Our conversation focussed a lot on the role and relevance of technology in (peace) negotiations spanning social media networks, how first interpretations of official processes are now made, remade, contested, contrasted, accepted and rejected, in close to real time, online, how instant messaging conversations go totally under the radar of the usual media monitoring (that informs an official process), the role of corporate entities and how 18-34 year olds, in various countries, engage with news and information.

Build Peace has a useful collection of tweets from the session here.

Frontier Issues: Some thoughts from 2012, still relevant

In early 2012, Patrick Meier emailed me (and I think a few others) asking the following questions: If you had some of the most cutting edge software developers at your disposal and funding were not an issue, what major software/computing innovations would have the greatest impact on disaster affected communities and humanitarian response? What are the most important gaps in humanitarian technology? What software challenges, if any, do you face in your own humanitarian work?

I rediscovered my email to him today, in light of some discussion I had in New York with the United Nations. Enough time has passed to publish what at the time was a bilateral exchange. I wonder, how many of these issues remain valid today, as those that will impact how we work in the years to come?


Machine translation and semantics and demonstrate how far even in a few years machine translation’s come. This is especially pertinent when non-English (script + language) data flows during most disasters (political or natural) eclipse the English you and I would be familiar with. During Strong Angel III we were shown a a real time translation of a TV broadcast – – but the technology still has a way to go before it can pick up nuances so vital for aid coordination during a crisis. I strongly feel however that NLP will play an increasing role in aid systems, and it appears the US govt (for parochial reasons) is getting into the act in a big way too –, along with the phenomenal work (which I believe you’ve covered in your own blog) of Recorded Future ( not so much for their platform, but their underlying analysis engine. And in the field of semantics, platforms like Cognition to Wolfram Alpha (powering Siri) are changing the way we interact with the web using strict Boolean logic. Can these new interfaces be applied to humanitarian platforms?

Not enough conversation is around the ethics of data generation, sharing, use and archival. and the more detailed (with my post tsunami experience) are early cracks, and along with more recent and in-depth writing on the use of Big Data (A brief exploration of Open and Big Data: From investigative journalism to humanitarian aid and peacebuilding), deals with this issue which I feel is often underplayed. Intertwined with issues of privacy, safety and security, the ethics governing the use of crowd-sourced information is put on hold for what are often called more immediate needs, but if unaddressed, can increase the risk of communities that were vulnerable. Lives saved during a disaster, ironically through the appropriation of information generated by them, could leave to lives lost to civil strife within repressive regimes.

Mapping are hugely interesting experiments, though field utility is suspect for at least 5 – 10 years. The whole gamut of physical sensors interacting with virtual design elements that influence data representation is a model of thinking that can however deeply inform humanitarian aid dashboard design and deployment. Obviously, Google’s Project Glass also holds promise at the field level for aid workers unfamiliar with the terrain. It is the most compelling vision to date of many other augmented reality platforms and apps already preset and working for Android and iOS. I actually started to talk about the use of augmented reality for humanitarian aid 6 years ago! See which I followed up in 2009 when Layar came to my notice I don’t know where Nokia’s at, my Layar has gone through many iterations.

Grassroots / citizen mapping
It may not be the case in every place and context, but essentially the technologies and tools for citizen mapping will grow. Products like will increasingly become hobbyist kits, complementing the kind of work done by grassroots mappers around the Gulf Oil Spill. Essentially, our view of the world is going to be increasingly plural – no one view will dominate another easily, with technology and tools to complement, confirm and contradict ground realities not just in the hands of govt’s, but in the hands of ordinary people too. The perceptions of Kibera to New York will change as a result, and this neo-geography will also inform identity – the sense of location within a society and community. From crisis to governance, these tools will play an increasing role.

Citizen journalism
In a blog post of yours from a while ago, you wanted a red button application for citizen journalism ( Now there’s one Of course, the Gulf Oil Spill resulted in a similar app – Along with the likes of Google + for iOS, FB Timeline, platforms like, and the really interesting, we are looking at the, interestingly, the fracturing of a key need – personal information curation. There are really complex algorithms behind each of these platforms and apps, and their potential to be deployed in and adopted for dealing with the peaks of information generation during a crisis are as yet untested.

Visualisation and mobiles
Can we do what does for books and music to missing persons registries and conflict drivers? Can we build remote field intelligence to mobiles so that what is shown on thin client apps in the field is geo-fenced, information rich, bandwidth frugal, contextual, updated, interactive and accountable? The combination of NFC, geo-positioning, data transmission via SMS, smart devices, multiplatform apps all exist – no one is really putting them together in the same eco-system, to create a HQ to remote aid worker ERM system of sorts. It can be done technically, but needs political vision and drive?

GAP: Archival, both the thinking and the tools
I first wrote about the problems of digital archiving in 2006 – The problem is growing. And fast. We’ve already lost, irrevocably, so much of the data produced during disaster even over the past 3 – 4 years. Given the pace at which information generation during and immediate after a disaster is increasing, the sheer technical challenges involved in archiving this information for posterity are significant, never mind the challenges over data governance, use and archival standards and media.

In 2006 I came up with six mantras – They remain valid today, and will I submit also be valuable into the future. And please, more Failfares – The marketing around specific platforms, apps and tools is already just to nauseating, because so much of it is disconnected from the more humbling ground realities. If we want a better future, let us start with our failures today.

Social media and government

Was invited by out-going Director General of the Government Information Department Ranga Kalansooriya late Sunday night to give a presentation the next day on social media and its importance for government officials, at a workshop organised by him with representatives from Facebook (coming in from India) present. Incoming / acting DG of the Government Information Department Sudharshana Gunawardene was also present for most of the workshop.

I used a few hastily created slides (download the deck here) to showcase why it was important government (officials) used social media, and Facebook in particular, to communicate to and if they so wished to do so, engage citizens. Noting the youth bulge in Sri Lanka’s demographic coupled with the rise of broadband and mobile phones, I flagged the sea-change in the way citizens got information and news, and indeed, went on to share this content with others. Knowing that Facebook representatives from India would go into it in more details, I briefly flagged the 5 million monthly active users on Facebook as reported on my ad-dashboard (which is actually much less than what Facebook itself said were Sri Lanka’s MAU at the time of writing this, which is 5.8 million). I also noted the overwhelming majority of these users connected through their mobile phones.

I then flagged to what degree Facebook played a role in Sri Lanka’s elections, just over 2015. I just showed three key slides, which suggested citizens were keen to know more about what their elected representatives did over social media, and also went on to inform those not on social media with news and information they had first learnt about online.

Screen Shot 2017-10-17 at 12.03.51 PM.jpg

Screen Shot 2017-10-17 at 12.03.57 PM.jpg

Screen Shot 2017-10-17 at 12.04.02 PM.jpg

Highlighting the substantial, in-depth research and polling on Facebook content and use in Sri Lanka done by the Centre for Policy Alternatives, I focussed on the ‘Consumption and Perceptions of Mainstream and Social Media in the Western Province‘ report, published early 2016 with fieldwork done late 2015. CPA’s website has more details about this report and key findings.

Showcasing a number of instances where the current President and Prime Minister were surrounded by youth posing for selfies, I noted that selfies were 2017’s autograph, and communicated a powerful political idea that the way to influence their minds was through the devices they owned and wanted to be framed through.

Screen Shot 2017-10-17 at 12.06.47 PM.jpg

Flagged a very interesting data-driven research paper on the use of social media by government official and public sector workers, I noted that it was not so much fear of social media that prevented better and more strategic use, but fear of reprimand, and the Establishment Code as it stands today, which is anathema to proactive disclosure and meaningful engagement with citizens.

The academic paper’s findings seemed to be borne out by those present at the workshop!


I also flagged the Information and Communication Technology Agency’s 2015 attempt to come up with a Social Media Policy for Government, under the then head Muhunthan Canagey. After two or three meetings at ICTA, and a process of eliciting input and ideas by placing the draft in the public domain, ICTA just forgot about the process completely.

Screen Shot 2017-10-17 at 12.10.48 PM.jpg

I suggested it’s revival, under or led by the Government Information Department, along with the recently released social media guidelines for New York Times staff, as foundations that could be adapted and adopted for the creation of a suitable framework for social media use in Sri Lanka’s public sector.

I ended by showing just how many from government were already on Twitter, and the hard road to social media’s meaningful use, from just a passive publication mode to active engagement.

Was also really good to engage with Facebook at the workshop, and learn what plans they had for the region and Sri Lanka. As noted on Twitter,

Rohan Samarajiva, who also spoke at the event, has a great post on what he said. Many important points there. The end  is particularly revealing. I was the 32nd to read the post on the official government news portal, in Sinhala. A day after the event was held.

By way of comparison, just the first of a chain of tweets I published on the event, late yesterday, had at the time of writing this generated 939 impressions and 45 engagements.

That’s social media’s power.


Hacking the referendum

This level of demographic targeting, increasingly possible even in a small country like Sri Lanka along with more conventional means of propaganda, circumnavigates the labour intensive electoral system and the problems therein of mass scale vote rigging. What could in the past be achieved by more traditional means of violence and intimidation can now, on a daily basis, be engineered by carefully crafting media content that spreads over social media, shifting, over time, entire groups against or for ideas, exploiting what endures as an information and media literacy deficit. Put another way, the explosive growth of social media is in fact a risk for progressive, democratic forces, because it provides easy, cost effective vectors through which spoilers can now influence and reach key demographic groups, who don’t go to political rallies, have multiple, liquid affiliations with mainstream politics, aren’t card carrying party political members and don’t engage with mainstream media through broadcast and newsprint. However, what is a risk is also an opportunity.

The government, despite strong opposition led by senior monk Anamaduwe Dhammadassi Thero, said last week that it will nevertheless continue to pursue the constitutional reform process, which will be put to a referendum. This comes after PM Wickremesinghe was reported in the media the week before saying the task of the Constitutional Assembly Steering Committee (CASC) will be to draft the new constitution in such a manner that will not require a public referendum. Adding to this confusion, data in the public domain over two years from Social Indicator, the social polling arm of the Centre for Policy Alternatives, suggests quite simply that if a referendum is held in the near future, the socio-political context is such that it is very difficult to see how a Yes vote would win. Given that a referendum is really an electoral litmus test of governance, entirely independent of the questions asked, the government…

View original post 1,297 more words