Ten years after the Asian tsunami, new ideas for DRR: Keynote address at ISCRAM 2014

I was invited to deliver a keynote address at the ISCRAM Asia Conference 2014 held in Sri Lanka from 20 – 21 June 2014. My session was held at the University of Colombo’s School of Computing and was titled ‘The more the technology, the more it stays the same?’.

Had my keynote been a week earlier, I would have begun in a different manner. As it happened, the keynote was during a week that saw some of the worst violence against the Muslim community in recent history, covered at the time not by mainstream media or official government sources, but by social media that didn’t exist at the time of the Boxing Day tsunami in 2004. I flagged the coverage over social media – from the start to the height and the immediate aftermath of the riots – of the violence in Aluthgama and Beruwela, noting that the social dynamics of information production, dissemination and close to real time correction, without any regulatory or oversight framework. I flagged the difficult for users new to or with average competence around Twitter and Facebook to ascertain verified information from the heightened production of updates around this time, not all of which were accurate or true. I did note that this information rich landscape particularly after a disaster was a feature of the emergency itself as well as emergency response, when ten years ago, all we had was POTS and in some cases, SMS, as information agents closer to real time.

Background

I framed my submissions in the keynote not as the expert I was (kindly) introduced as, but as a student of social media by virtue of curating Groundviews (including its Facebook and Twitter accounts) for over eight years. I noted that because of the curated nature of Groundviews, I had read or produced over eight years (up until the time of the keynote, and not including content on Facebook and Google+),

• 4.6 million words on Twitter from over 32,900 tweets
• ~7.3 million words over 49,700 comments
• ~5.8 million words of content over 2,900 articles

This I said gave me what is almost muscle memory when navigating social media – the ability to see dynamics, intuit flows, verify information, curate content and help shape the information landscape around vital debates, including in the case of Aluthgama, political and social disasters.

Developments since 2004

I recalled how in September 2007, a tsunami warning (three years after the Boxing Day tsunami) created mass confusion and panic, something I’ve written on in some detail on this blog. The blog post goes into how information around the alert was distributed over SMS, and looking back, what’s interesting is that there was no impact of web based social media at all in the dissemination of critical alerts, or a high prevalence of citizen generated content around the alert with for example geo-located situation reports.

I then recalled the tsunami warning in April 2012, which again I’ve written on in some detail (Sri Lanka’s tsunami warning on 11 April 2012: Twitter and social media role). Noting that I was at lunch in Colombo when the alert was first received, I traced how even though mainstream media wasn’t plugged into social media at the time, how alive Twitter in particular was to the tsunami warning. I noted how various mobile voice and data failed in a cascading manner, linked to not just call congestion, but the inability of mobile networks to handle the heightened generation of data. I flagged how over Twitter, there was more timely, crowd verified information around the initial tsunami warning.

Looking at the dynamic generation of information (over social media in particular) and social network dynamics around this information, I flagged the importance of reading Social Physics by Alex Pentland as a way of understanding how today, information over social media impacts disaster response. In particular, I noted how on Twitter and very quickly, markers of trust and as a result, sources of verified information were very different to mainstream media personalities, institutions and official channels of information (e.g. the Meteorological Department).

Recent history

I then focussed on the tragic violence in Aluthgama and Beruwela, noting that social media had moved from the margins to being centre and forward in helping a wider public understand what happened. I noted how there was a rapid self-organisation of many on Twitter around the #aluthgama and #beruwela hashtags, how a few journalists and a number of civic media accounts (including @groundviews curated by me) become very quickly the predominant sources of information around the on-going violence, how the social networks were in the aggregate more information than mainstream media and official sources (a somewhat unfair comparison since both were completely silent over the violence at its peak), that even those who tried to spread false rumours and incorrect photographers were quickly named and shamed by the community, and how the government’s first and for a long time, only response to the riots, over Twitter, was a result of social media pressure. I also flagged the dynamic of what I called the social acceptance incentive – that at a time like Aluthgama’s violence, few wanted to be associated with, in their name, content that was erroneous, inaccurate or downright false.

I then looked at what had changed from Boxing Day 2004 to the week I did the keynote, in terms of the evolution of technology to deal with the increase in natural disasters and political emergencies.

I noted that in Aluthgama, information and communications technologies helped in securing information for a better, wider public record, in the absence of any official statement from the government and almost blanket censorship by the mainstream media. But then I flagged the example of when in June 2013, over 39 fishermen died due to lack of early warning around severe weather conditions, including many more missing and hundreds of buildings damaged. Saying that this wasn’t a one-off, I then flagged an example from November 2011 when 14 fishermen died, with many bodies never recovered, and by some accounts over 10,000 buildings were damaged, on account of a storm. Here too, there was no early warning and worse, as reports claim, information around the possibility of severe weather conditions held with one government agency never transmitted in time to the agency responsible for disseminating this early warning. I asked how technology could help strengthen early warning and disaster response when there was clearly a lack of institutional investments (both human and financial), and a politicised bureaucracy.

Noise vs. actionable information

In terms of information, I noted how greater, faster more responsive architectures – created and supported by in large part mobile telephony’s growth as well as coast to coast broadband in Sri Lanka – did not mean that communities on the ground still knew what to do around false information, or partially accurate information. Rumours, ironically, in the age of social media, put at risk far more than even just a few years ago.

I also flagged that despite the advances in technology, there was still a bias towards attention around sudden onset disasters (e.g. floods and earthquakes) over longer term complex political emergencies (like CAR, South Sudan or Syria). Given that there are hundreds of Aluthgama’s in a single day in for example Aleppo, what could I asked technology in general do to identify and transform the violence, if only in the humanitarian domain?

Information rights and ethics

Rights and privacy are two issues I’ve been very passionate on, and I asked to what degree those involved in innovating technology around disasters had thought of them, especially considering that passive or active participation in information network could potential put at greater risk individuals or communities who data was part used first for first response, but then were in the public domain without any regulations or legal frameworks around reuse. From the governance side, I focussed on the technical side, flagging that for so many of the leading platforms around disaster response and early warning, systemic interoperability was still not possible. I flagged the HDX project by UN’s OCHA, and the HXL standard that’s part of it, as a potential technology to incorporate into these platforms.

I looked broadly at how technology had developed at an increased pace after the Haiti earthquake in 2010. The development of the volunteer and technical community, crisis mapping on the web, the digital humanitarian network, the use of UAVs for humanitarian work, neo-cartography and OSM progress on the web, platforms like Tomnod, participatory mapping and grassroots mapping initiatives, data visualisation platforms, social media channels are all developments I said had very rapidly changed how institutions like the UN looked at and approached DRR, early warning and disaster response. I said that though the emancipatory power of technology was high, innovation and development were still hostage to culture, politics, experience, history, memory and governance. I also inquired as to why so few at national government level were talking about hackathons or interactions with the tech community around disaster resilience innovation, recognising the fact that the first responders have always been those who have also been affected by the disaster.

In light of the tsunami of information around a particular disaster, I said that many today feel they are better informed but in light of Eli Pariser notes, are actually hostage to filter bubbles, where they are ignorant of competing or complementary information that can impact decision making.

Ideas
I ended by keynote by leaving the audience with some key ideas. How for example a progressive government could work with civil society and the private sector to strengthen DRR, early warning and rapid response. I noted that instead of waiting for emergencies, this interaction could take the form of a sustained dialogue, looking at for example crowdsourcing or using ICTs to look at deforestation issues, the rise of urbanisation and the need to create disaster resilient urban spaces, climate change impact, and the safety of artificial water bodies like tank bunds.

Leverage telecoms client base

• Using the bills posted or emailed to every single one of the millions of subscribers to mobile or fixed telephony services across the country, or using the the utility company client records (e.g. electricity and water), questions could be asked around raising awareness on DRR, preparedness, evacuation or response to disasters by small questionnaires or information boxes in the monthly bills.
• In the event of a disaster, consumers could be told to give micro-donations (Rs. 5) coupled with a matching sum by the telecoms or utility company, from their phone and over USSD or SMS, which could go towards aid efforts.
• Use the printed bills to raise awareness around dengue prevention, and with the addition of MCQ’s, match correct answers with a percentage point reduction in the total bill cost. One could also gamily this to publish on social media those who repeatedly scored the highest points, thereby creating more interest around the questions.

Leverage e-commerce sites

• Use the user logins to or checkout verifications on leading e-commerce sites like Takas, Wow.lk or Kapruka to get registered users to answer questions around, for example, the conditions of a bund around a tank. This would function like a captcha, but work more like Tomnod, where large sets of imagery would be distributed amongst millions of subscribers or consumers, so as to crowdsource what would other be an impossibly labour intensive job.

Leverage loyalty points

• Have a single portal from which loyalty points from various cards and services (Nexus, HSBC credit cards etc) could contribute towards disaster relief efforts, allowing one to donate points from whichever service as cash.

Take Sahana to schools

• The week saw a lot of developers talking about and working on the latest version of Sahana. I asked why this could not be taken to schools, to help students at a young age – through guided classes and interactive simulations around disasters, using Sahana, learn about DRR and importantly, the use of information architectures in disaster early warning and response.

The Q&A session focussed, unsurprisingly, on social media’s role in and around the riots in Aluthgama, but there were other questions too – ranging from what impact had been made to encourage telcos in particular to focus on disaster relief and post-disaster technology frameworks, what real impact had been made in policymaking around disasters and technology, use cases and example of ICTs and social media to prevent or mitigate conflict (I flagged Umati as one example) and more generally around the ideas I mooted around how, essentially, clients, consumers or citizens can donate time or money – through established logistics, billing or merchant networks – to support relief work as well as in crowdsourcing efforts around early warning and DRR.