Could ‘Crowdchecking’ Help Manage Disasters More Effectively?

By August 20, 2014
Verily crowdchecking

The emergency services in many countries are keen to make more use of social media in order to help manage crisis situations. UK researchers are now studying ways in which what is being termed ‘crowdchecking’ might be used to good effect.

According to an a posteriori analysis of the tornados which struck the United States in 2011 by researcher Thomas Niederkrotenthaler from the Centre for Public Health at the Medical University of Vienna, information from social media is generally regarded as more reliable than the broadcasts on official media channels. Nevertheless, this information still needs checking. With the aim of helping the emergency services respond more effectively to humanitarian needs in disaster situations, researchers at the University of Southampton in the UK have just taken part in the development of Verily, an Internet app designed to collect proof of information posted online during natural disasters. Given that information on the social networks is usually simply shared without undergoing any official verification, the organisations responsible for providing emergency assistance frequently find themselves being fed contradictory information.

Testing the responsiveness and effectiveness of fact-checking alternatives

The Verily tool is part of the current trend towards fact-checking. It has been designed to ensure that reliable information can be obtained as quickly as possible in a crisis situation. Looking at the example developed to test the viability of the method, we see that it relies on verification backed by ‘justification’ – i.e. proof – agreed by the Internet ‘crowd’. For example, one of the questions in the test package – the #VerilyLive Challenge – asks people to confirm that the street in a certain photograph is in Rome. One of the app beta-testers answered ‘no’, proving the negative answer by relying on another photo in the list which was taken in a different town in Italy. The range of techniques used to fact-check the answers proved to be very effective. However, it remains to be seen whether the model would be equally as effective in a real crisis situation.

‘Crowdsourcing’ information to help the emergency services

A number of tools have already been developed to detect posts – out of all those published on the social media – which could be useful to the emergency services in times of crisis. One example is Twitcident, developed by Dutch company CrowdSense, which is based on artificial intelligence. Crowdsourcing of information – which demonstrates its great popularity every day of the week on websites such as Wikipedia and TripAdvisor – is now looking highly promising for use in natural disasters. If the ‘crowdchecking’ process proves reliable, it would enable the emergency services to make better use of the information posted on social media. Everyone recognises the value of crowd sources and agrees that they can be very effective, but many people still feel that, due to the general uncertainty that surrounds these unofficial sources they cannot be used systematically as a basis for decision-making.

Legal mentions © L’Atelier BNP Paribas