In 2013, Wayan Vota, now a developer at nonprofit human-development organization FHI 360, argued that humanitarian organizations should treat "communications as aid." Vota's argument reflects a growing consensus that "humanitarian technology" is an important resource for improving disaster planning, understanding conditions on the ground, and coordinating delivery of aid and that restoring communications networks in areas struck by natural disasters is now as important as delivering food and emergency medical care. Although it has mainly been a tool for international organizations and relief workers, humanitarian technology in the future will also help improve communities' ability to prepare for and recover from disasters.

Disaster Planning and Preparation

Pervasive computing's influence on disaster response begins with emergency planning and preparation. Traditional disaster preparation and planning relies heavily on tabletop exercises. Most tabletop exercises take place around a large model of a city, airport, or other physical space; a facilitator directs the action, announces new challenges or surprises, and leads after-action reviews.

Tabletop exercises help responders visualize how emergencies might unfold, see their own actions as part of a whole, rehearse their responses to scenarios in the company of other first responders, and learn to improvise on the basis of formal plans. Whereas scenario participants often make use of sensors and communications technology both to become familiar with these tools and to identify incompatibilities between systems used by different responders, researchers are now developing tools to support learning within scenarios. A group centered at Ireland's Waterford Institute of Technology, for example, is working on "Cooperating Smart Spaces" that mix wearable devices and sensors, large-scale displays, and collaboration tools. Researchers in the Universidad Carlos III de Madrid's Interactive Systems Lab have developed a proof-of-concept system—SharedViews—to support development and refinement of emergency plans. A distributed-computing environment including mobile-app clients and cloud-services integration lets teams (and, in theory, members of the public) collaborate on emergency-planning processes and integrate lessons from scenario exercises.

Disaster Mapping

Disaster mapping is the area in which humanitarian technology has scored its most notable success. Since 2004, GIS professionals have been developing participatory mapping systems for use in emergencies. Today these mapping projects fall broadly into two types. One uses community efforts to create maps of villages, new towns, and other areas that official bodies have ignored. The OpenStreetMap project (www.openstreetmap.org) aims to create open-source maps as an alternative to both government maps and corporate GIS projects like Google Maps. For rapidly growing cities in the developing world, these maps may be the most reliable and up-to-date maps available. The Humanitarian OpenStreetMap (http://hot.openstreetmap.org) overlays these base maps with photos and information from text messages, tweets, and other sources to give first responders a clearer picture of conditions in disaster areas.

During the 2010 Haitian earthquake, Humanitarian OpenStreetMap produced the most reliable and fastest-updated maps of affected areas; not only were they used by relief workers, they were even used by the US military. This success led to the creation of Crisis Mappers Standby Task Force (SBTF), a volunteer network that can mobilize quickly to create maps immediately after disasters. During 2010 and 2011, SBTF created a number of maps for the United Nations Office for the Coordination of Humanitarian Assistance (OCHA), including maps of Chile after the 27 February 2010 earthquake, of Pakistan after the July 2010 floods, and of Libya during the 2011 civil war. In April 2012, OCHA helped create the Digital Humanitarian Network. Because they are distributed and include hundreds of volunteers, these groups can produce and update maps far more quickly than can aid agencies. SBTF and other volunteer groups collaborated to produce disaster maps of the Philippines 12 hours after Typhoon Bopha (sometimes, Typhoon Pablo) struck in December 2013. Further, the maps themselves serve as a useful filter and interface for aid workers. They give workers a familiar, visual way to access photographs, social media, and other data, while excluding information that is not immediately relevant.

Humanitarian-Needs Prediction

The 2010 Haitian earthquake was also the first in which researchers tried to use social-media data to forecast how people would react to the disaster. The earthquake severely damaged Haiti's most populous city, Port-au-Prince; it killed or injured half a million people, and destroyed or damaged 300 000 homes. The aftermath of the earthquake forced 1.5 million people into temporary housing and camps. A team of researchers based at Karolinska Institutet in Stockholm, Sweden, worked with Haitian cellular service Digicel to analyze anonymized records of 2 million mobile users to predict where they would relocate. "When disaster strikes we tend to seek comfort in our nearest and dearest," according to Karolinska Institutet researcher Xin Lu in a June 2013 public announcement. The data showed that "where people were over Christmas and New Year, which was just before the earthquake, tended to be the place where they returned." Some 600 000 people fled Port-au-Prince after the earthquake, but most returned after a couple of weeks because the countryside couldn't support them, and most relief supplies were in Port-au-Prince. Had NGOs had the Karolinska Institutet forecasts, they could have planned to deliver supplies to smaller cities where people had sought refuge.

Although the work was too late to be of use to aid workers organizing refugee camps in Haiti, the Karolinska Institutet's work proved very useful months later, when people began fleeing a cholera outbreak in St. Marc. The researchers worked with Columbia University epidemiologist Richard Garfield to forecast where they would go and were able to alert aid workers and public-health officials in other cities to prepare for an influx of refugees. The group has since formed a nonprofit research group—Flowminder (www.flowminder.org)—to generate forecasts for other disaster areas.

Augmented Humanitarian Technology

As the number of relief organizations operating in disaster areas has increased, and as aid workers and victims both acquire the means to collect and broadcast information during disasters, humanitarian groups face a new problem: verification and management of large quantities of data, while also coordinating relief and supply efforts. Disasters now play out in real time in blogs and social media: The Haiti earthquake in 2010 was the first in which relief organizations solicited funds through Twitter and other social media, and participants themselves reported using those same channels to report about conditions on the ground. But as social media increase the visibility of disasters and attract more relief, they can also lead to bottlenecks and confusion.

The Orchid Programme, a collaborative academic-industrial project headquartered at the University of Southampton, believes it can develop artificial-intelligence tools to improve disaster response. (According to the project's website, Orchid is not an acronym but "a metaphor for how a system can flourish"—as certain orchid plants both grow on trees and are beneficial to them.) Orchid aims to merge human and artificial intelligence into an efficient complementary unit—a Human Agent Collective (HAC). According to a January 2014 SciDev.Net article, HACs will be active in "planning the flight paths of surveillance drones, verifying the authenticity of information coming in from social media, facilitating data sharing and organising human teams based on their skill sets and current needs on the ground." (Drones and other unmanned aerial vehicles are not yet common in relief work, but enthusiasm among volunteer technical experts, in combination with efforts by the drone industry to market drones as humanitarian technology, strongly suggest that they will serve in disaster relief in the future.)

One of the major challenges Orchid faces is in making HACs that can function in an environment in which workers must improvise, deal with voluminous yet uncertain information, and rearrange their priorities as situations change. Traditionally, the design of automated systems in the workplace has emphasized the use of formal procedures to focus workers and maximize efficiency, creating what sociologist Aneesh Aneesh called "algocratic" modes of organization. Orchid researchers recognize that their systems will operate in information-rich yet chaotic environments, with resources and personnel shifting from one area or task to another. HACs therefore must be able to adjust their priorities and collect and present different types of information as tasks change. However, if Orchid succeeds in developing HACs that support flexible work and quickly shifting teams that learn to provide relief workers with relevant information as their needs change, this success would be an important advance in computer science that many other workplaces could use.

The Future of Humanitarian Technology

Ultimately, though, the real breakthrough in humanitarian technology may not come from helping international agencies and NGOs work more effectively but from giving affected communities tools that enhance their ability to respond to disasters. Although Western media tend to concentrate on the roles of foreign relief workers in disaster response, only about 10% of people affected by emergencies actually receive direct help from relief agencies; the rest are saved by neighbors, friends, and family and rebuild their lives with the help of local institutions and diaspora networks. Consequently, some people argue, humanitarian technology should reverse the current balance of power between relief organizations and the people they serve. As Toby Porter, emergencies director for Save the Children, indicated in a 2007 article in the Economist, such a model would turn international agencies into the equivalent of Amazon.com, delivering supplies and services ordered by local communities.

Such an approach could be problematic during some crises: Famines can be a consequence of civil war, and relief agencies already have to deal with politics and corruption. Humanitarian programs aim to empower ordinary people, not the most unscrupulous, media-savvy players. Nonetheless, international-development-technology specialist Wayan Vota argued in 2013 that the time has come to move away from thinking of recipients of aid as passive "beneficiaries." New technologies enable them to act as "customers, clients, co-creators, or...constituents"— as active participants who drive the aid and reconstruction process. If designed well, tools that allow affected people to plan and direct their own relief, to work with aid agencies, and to track the delivery of supplies would empower the people who save the most lives, better serve long-term development goals, and allow international agencies to serve larger numbers of people more effectively.