Crowdsourcing, the practice of soliciting ideas from large groups of people, is a rapid, inclusive and effective way for humanitarian responders to gather information in the immediate aftermath of a calamity.
Consider the example of Ushahidi, an open-source software used to collect information from local people and present it in the form of interactive maps. Created by a group of young Africans during the 2007 post-election crisis in Kenya, Ushahidi has been used farther afield – with differing results.
When Ushahidi was used in Haiti following the 2010 earthquake, it proved effective in providing information about the needs of affected communities – despite the challenges of verifying the accuracy of the crowdsourced information. In a mostly urban environment, where communication networks were quickly repaired and mobile phones were the main means of communication for survivors, people from affected communities provided eyewitness reports through short message service (SMS) and social media. An international team of volunteers compiled the information and displayed it in real time on an online map that could be accessed by anyone, including responders.
But when the software was used in the Democratic Republic of the Congo, in 2008, to map incidents of violence against people who had been uprooted from their homes, the local population did not embrace it. Challenges included the lack of access to technology in remote areas, language barriers and security concerns. The likelihood of false information and corruption was compounded by the challenge of verifying the factual accuracy of eyewitness reports in regions without networks of trusted sources.
Where access to technology is skewed by wealth or other manifestations of power, crowdsourcing will reflect and can magnify existing inequalities. For example, a 2013 Gallup poll of mobile phone users from 28 countries in sub-Saharan Africa found that mobile phone ownership is lower in rural areas and among poor households. This means that the needs of poor families and rural populations are likely to be left out of crowdsourced information – as reports delivered through SMS could merely reflect the needs of those who own mobile phones.
It can also be difficult to distinguish between useful information and false or irrelevant data. This can be a particular problem when lots of information comes in during a short period of time – as happened after the 2010 earthquake in Chile, when 7,000 tweets containing information about the event were sent in the space of one hour.
Algorithms that distil reliable information from the deluge of data are available and can be effective in helping responders make sense of large amounts of information in a short amount of time. Even so, it remains difficult to react rapidly in an emergency based on crowdsourced information alone. Crowdsourcing can contribute to awareness of a situation – not least by offering a barometer of public sentiment and an additional means of corroborating facts and filling gaps in existing reports. But tried and tested means of gathering information, such as coordination with government authorities and consultations with stakeholders, remain essential to any attempt to respond to an emergency.