Monika Büscher and Michael LieglABSTRACT
Mobilities.Lab, Centre for Mobilities Research
Department of Sociology, Lancaster University
Social media enable new practices of public engagement in formal emergency response, but the scale and speed of innovation at this juncture has outpaced the development of ethically, legally and socially ‘virtuous’ practices and technologies. We discuss positive and negative frictions and avenues for socio-technical innovation that bridge between connected communities in crises and professional responders.
Almost every post-disaster report highlights a need for better integration of affected communities in emergency management. The US Federal Emergency Management Agency (FEMA), for example, argues that natural or manmade crises (floods, storms, violent attacks) can be addressed better with a ‘Whole Community’ approach, where ‘officials can collectively understand and assess the needs of their respective communities’ and communities can play an active part in emergency planning and management . In some sense, this acknowledges communities as an agency in multi-agency crisis management. However, it is practically and politically difficult to switch from approaches focused on protecting and controlling the public to engaging with communities. This is exacerbated by the fact that the notion of a clearly defined community whose needs can be assessed by ‘their’ respective officials is flawed. Innovation in ‘crisis informatics’  has made social media an important force in connecting people within and beyond local communities. This has begun ‘to fundamentally alter the very nature and arc of emergencies’, bringing disruptive innovation, that is, innovation that changes existing social, material and economic practices , to crisis management. This paper explores ethical, legal and social challenges and opportunities at this juncture. After a review of background research (Section 2), we examine a selection of examples (3). In Section 4, we discuss positive and negative frictions, and we conclude by delineating some promising avenues for socio-technical innovation (5).
Local communities can be connected through complex communicative networks and in crises extend links to national and global communities, including diasporas and crowds of volunteers and emergent ‘digital humanitarian organizations’ who perform increasingly official responsibilities of gathering, verifying, geo-locating and mapping information from afar (such as CrisisMappers, StandBy TaskForce (SBTF), Virtual Operations Support Teams (VOST)). This can allow faster and more detailed awareness of the needs of affected communities and the nature and extent of damage, which makes crisis informatics interesting as an informational service for official emergency responders.
But crisis informatics can also service practical self-organised mobilisation and coordination of local resources, knowledge, and efforts. During the floods in Germany in 2013, for example, 29% of Twitter-messages focused on coordinating help and resources locally . Reports from sandbag filling stations appeared alongside calls for help and a crowdsourced map of the current need for volunteers in different places . Lüge  suggests that these examples index a shift in the use of social media for emergency management. The informational service function for official response is increasingly seriously complemented by a practical service function for self-organized community help and resources, especially for members of the public. Yet, research in crisis informatics is still mainly focused on understanding and developing means of extracting more valuable, reliable, ‘actionable’ information from social media for enhanced situation awareness, especially for professional responders. There are a small number of studies that explore how members of the public self-organise disaster response in situ through social media and opportunities and challenges that arise around dovetailing these efforts with formal response operations, for example, after the 2010 Haiti earthquake , or the 2011 ‘Super Outbreak’ tornado in Alabama . They raise many questions, and amongst them important questions about ethical, legal and social challenges.
Regarding the ‘curation’ of crowdsourced information for situation awareness , practitioners and researchers already analyse and address ethical, legal and social issues, ranging from misinformation through information that can undermine ‘information superiority’ and endanger operations  , to vigilantism , tort liability (for civil wrongs) for volunteers and various challenges for professional responders . There are calls for a ‘code of ethics’  and some early formulations of ‘Twitter Commandments’ for ‘voluntweeters’, providing ‘guidance about sorting accurate from inaccurate rumor, and for “tweeting responsibly” during disasters’ , as well as guidelines for crowdsourcing information from populations affected by conflict . However, these focus on practices of information extraction and processing, and less on information use for self-organised mobilization of resources by nested digital and local communities. But ethical, legal and social issues arising at the interface of such community efforts and ‘Whole Community’ approaches to crisis management should be part of discussions about ethical codes of conduct and the practical ethics of such activities. And the challenge does not end there. A mere code of conduct is not enough. Ethical principles must be enacted as moral conduct and – when such conduct is digitally augmented – this also calls for technologies that actively support virtuous moral practices.
To better understand digitally augmented emergency response practices and ethical, legal and social issues, we now examine a selection of examples. The focus is on the interface between self-organised community efforts and official efforts during the response phase of crisis management.
A. The Ushahidi Haiti Project
January 2014 was the 4th anniversary of the Haiti earthquake, where over 220,000 people were killed and over 300,000 were injured. The earthquake made more than 1.5 million people homeless, and resulted in an ‘immense humanitarian crisis, highlighting long-lasting development challenges’ . The Ushahidi Haiti Project (UHP) marks a milestone in the development of crisis informatics for humanitarian emergency response. It provided novel ways of crowdsourcing and mapping information, even supporting the task of deploying resources to people in need. Morrow, Mock, Papendieck & Kocmich  for example, describe how the Department of State Analysts for the US government interagency task force and US marines used UHP information to enhance situation awareness and identify “centers of gravity” for deployment of field teams (ibid, p.4). Innovations like ‘Tweak the Tweet’ (TtT), a standard which suggests a uniform format for reports through hashtagging needs, locations and contact details, promoted a shared ‘grammar’ that facilitated computational parsing and mapping of tweeted information . Starbird & Palen observe how volunteer translators or ‘voluntweeters’ translated reports from different sources, such as text messages or tweets, using the TtT syntax in response to the Haiti crisis, and worked as ‘remote operators’ to facilitate assistance, resource coordination and collaboration from a distance. Amongst other things, they promoted the international transfer of small funds via Paypal to many Haitians’ pay-as-you-go mobile phones, and even coordinated the provision of trucks to specific locations and local volunteers, with messages sent back and forth and including confirmation of resolution of resource coordination challenges. However, despite the successes Morrow et al.  note, they also found significant barriers to the use of microblogging by official responder agencies. They quote one of their most experienced emergency responder interviewees as describing UHP as ‘a shadow operation that was not part of the emergency response plan’. Further to this, Starbird and Palen  describe how voluntweeters felt frustrated and ‘obstructed when the “formal” response moved into place’. One of the most challenging issues for integrating local, online and official response was the reliability of information. Despite the fact that many experts and government organisations like FEMA, the Department of State as well as international organisations like UNOCHA agree that an integration of digital humanitarian organisations (DHOs) with formal emergency response efforts is invaluable, and while they are establishing interfaces to grassroots networks, there are some serious obstacles:
Federal agencies are legally obligated to provide data that are accurate, reliable and useful. They must take steps to ensure the integrity of information, … prevent the release of data that breach the privacy or security of citizens or organizations, violate nondisclosure agreements, or endanger national security 
These constraints are hard to overcome, and they are accompanied by operational challenges and opportunities.
B. The Norway attacks
The analysis in this example builds on previous work we have published with colleagues in . Names and the wording of twitter messages have been altered (maintaining the sense of the contribution) to anonymize persons.
On 22 July 2011, two attacks took place in Norway, a country that ranks second in Europe for internet use , and the local and digital communities’ response to the events highlights important issues in relation to operational synergies and frictions between community and formal response efforts. The first attack was a bomb explosion in the government quarter in Oslo. It killed 8 people and wounded 30. Less than two hours later, another attack took place on the island of Utøya, about 40 km northwest of Oslo, at the Norwegian Labour Party’s annual youth summer camp. The same person who had exploded the bomb in Oslo came to the island disguised as a police officer and shot 69 people dead and wounded 60 more. In a study of social media use during the response to these incidents, we found direct communications between people affected and a digital community converging to help by, for example, coordinating the opening of private WiFi in Oslo, and finding and sharing information about hospitals to call to find out about missing relatives . Then, shortly after the shooting on Utøya began, ‘NilsPetter’ received this tweet from one of the people on the island:
@cpltee @NilsPetter We are sitting by the lake. A man dressed in police uniform is shooting. Help us regarding when the police will arrive. [5:58 PM 22.7.2011 from Twitter for iphone]
At the point this tweet was sent, the police, who led the formal response effort, had already arrived at the mainland ferry pier, less than 1 mile from the landing pier on the island across the lake. However, there was concern over a second car bomb at the mainland ferry port and fears that there could be more than one shooter. By following NilsPetter’s stream after the request above, we learn more details about the shooting. Updates from the island received and broadcast in this way supported some insight into the events for twitter followers and their social networks, but they also caused great concern, especially amongst friends and parents, whose first reaction was often to find out about the whereabouts and safety of their friends or children. NilsPetter’s messages reveal that this caused problems. With some urgency, he tweets ‘DON’T CALL friends on Utøya’, explaining that ‘It can endanger them’ in a second tweet. The sound of mobile phones ringing could put them in danger, a fact that may seem obvious with hindsight but could be missed in the heat of the moment.
Further to these coordinating and awareness configuring communications, people in the surrounding area heard gunfire and saw people waving and calling for help. Some of them used their boats to pick up people from the waters surrounding the island. Numerous tweets and retweets encouraged this:
RT @elisefang: do you have a boat close to #Utøya? Pick up swimming children around Utøya! #osloexpl #norwayterror # bombeoslo [via Twitter]
It is difficult to establish whether members of the public helped young people in the lake specifically because of such social media messages. Unlike in Haiti, these calls for resources were non-specific in terms of the precise location where boats were needed and we found no messages that confirmed details or completion of private rescue missions. However, the fact that requests for boats were made in large numbers and over different channels suggests that emergent practices of ‘micro-coordination’ in public emergency response are taking shape. Ling & Yttri  coin this term to analyse everyday practices of flexibly coordinating everyday activities like shopping or meetings on the move, using mobile technologies. They describe this as micro-coordination, because dynamic and delicate coordination of people, places, times and objects becomes possible. Translating such social innovation from everyday contexts to crisis situations has the potential of supporting efforts of mobilizing needed resources (such as WiFi or boats) locally and swiftly. However, undertaking such efforts ‘in the wild’ also creates tensions and potential problems. The professional responders in Norway found themselves under intense media and social media scrutiny and pressure, and, in a situation where it was deemed too dangerous to allow professional emergency staff near the island, official emergency responders were seen to allow members of the public to rescue victims. Some of these victims were taken to hospitals without being registered, making it difficult to know where they had gone, what treatments they got, or what they still needed.
C. The Boston Marathon Bombing
Our final example brings out even more challenging issues. The Boston Marathon came to a sudden end on April 15, 2013 when two bombs exploded close to the finish line, killing three people and injuring an estimated 264 others . Within hours, the FBI called upon bystanders to submit their images and videos from the event, triggering a massive ‘crowdsourced intelligence gathering’ . Two days later the police released a photo of one of the suspects and asked the public for help in identifying him. But within these two days, ‘digital bystanders’ had not waited patiently. They had turned to ‘crowdsourced crime solving’ , analyzing image content, collecting clues and listening to and posting recordings from the police scanner. This was largely organized on social news and activism websites, ‘Reddit’ and ‘4chn’. When a tweet noted a resemblance between the suspect on the police photo and a tweeter’s former classmate, his name was posted on Reddit along with another name from the police scanner. This resulted in this widely retweeted tweet:
@KallMeKG: BPD scanner has identified the names: Suspect 1: Mike Mulugeta Suspect 2: Sunil Tripathi. #Boston #MIT . [April 19, 2:50pm, cited in ]
For a short time the crowd detectives celebrated a victory: ‘Reddit solved the bombing. Before the Feds’ . But soon the FBI and news outlets released the Tarnaev brothers’ names, exposing the crowd as ‘digital vigilantes’ who had spread rumours slandering two innocent men .
The crowdsourced manhunt after the Boston bombing highlights some of the risks officials take when collaborating with volunteers. It also showed that there is good reason for the media to be more cautious of using crowdsourced websites as a source . For unlike in official emergency and news agencies, the social media crowd is unorganized, untrained, uncertified and largely anonymous. There are some effective social informational practices of self-regulation in collective intelligence, but these do not seem to function in ‘manhunt’ circumstances . While there are some promising attempts to design automated support for detecting ‘signatures’ of misinformation , at this point it seems doubtful whether legal and institutional fears about trusting information from social media can be overcome anytime soon.
The use of social media for self-organised mobilization of resources, knowledge and efforts by nested digital and local communities raises opportunities for positively disruptive innovation in emergency response as well as challenges. On the one hand, the turn to crowdsourcing information and support for micro-coordination and collective intelligence to augment local communities’ capacity for self-help can help address needs swiftly and effectively. In the emergent ‘new reality’ for emergency services , such enhanced community resilience presents new ethical, legal and social openings. As increased frequency and severity of disasters combine with heightened vulnerability through ageing infrastructures and populations, emergency services are under pressure. Some see the future of emergency response in spreading the burden of responsibility, including to communities . On the other hand, allowing the practical service function of social media for self-organized community help to grow generates conflicts and frictions.
The examples above exhibit the momentum of social and technical innovation in crisis informatics and highlight different dimensions of how new technology and related participatory practices have mobilized emergency response, introducing new forms of agency and actors and provoking negotiation and contestation of competences and responsibilities. In this emerging new reality of emergency response we see three types of entities/agencies negotiating their relationships and roles:
In the first large scale mobilization of connected communities in the aftermath of the Haiti earthquake, distributed crowd communities were able to map affected areas and thereby provide a baseline for translating and mapping needs. However, who can address all these needs and how? Locating and making more needs visible may seriously exceed the capacity of formal response organizations to address them and make self-help a necessity. This opens up political questions over the resourcing of crisis management and emergency response. Further questions include: How can the reliability of the information provided by volunteer services such as the UHP be ensured? And what efforts are being made to bridge digital divides? People who do not have access or the skill to use digital devices may not be able to put themselves and their needs on the map. And finally, when do the responsibilities of the distant community members wane? 4 years after the earthquake, the United Nations find that 817,000 Haitians still need humanitarian assistance . Yet, most digital humanitarians have moved on to the next crisis.
The Norway example also shows how social media can be used to mobilise local resources such as WiFi and boats, to inform communities and ‘configure’ their awareness , and to support the channelling of enquiries to official helplines. But the example also showed how connected communities can expose needs that formal responders may not be able to address safely and place professional responders under immense and more immediate public and media scrutiny than was possible before the advent of social media, leaving them no time or space to think. In a context where formal emergency responders are not participating in the ‘twitterverse’, communities, in turn, do not have much access (outside conventional media) to professional reasoning about the unfolding complexities, or the professionals’ rationale for caution. But communities and societies have no right to expect emergency responders to be heroes, and they have a duty to fulfil their side of a social contract that provides emergency responders with the necessary means, respect and protection to carry out their work. Members of the public becoming involved in improvised rescue missions may lack situation awareness and may unwittingly put responders and victims at risk and create complications, for example, with a view to keeping track of victims and informing relatives.
The Boston example highlights a need to distinguish more carefully between the different types of emergencies that practical service functions of social media for coordinating connected community responses can be brought to. Equipped with images sourced from the ephemeral local community of visitors and participants in the marathon, an untrained, uncertified and anonymous crowd launched into crowdsourced crime solving and falsely accused two innocent men. The dynamic of such ‘clicktivism’ is powerful, feeding on and feeding into sensationalist media reporting in a way that raises questions about responsible social media use on behalf of all parties involved. But it also opens up questions over how self-regulating mechanisms that operate in other forms of connected community social media use in crises may be leveraged. Starbird et al.  explore how indigenous mechanisms of identifying and correcting misinformation may be computationally supported through automatically detecting corrections. This highlights the entanglement of social and technical innovation at the frontiers of crisis informatics and its transformational implications for the relationships between established responder organizations, digital volunteers and self-organizing connected communities.
The positive and negative disruptive innovations at this juncture call for creative, concerted efforts to develop ethically, legally and socially ‘virtuous’ practices and technologies. Automatic detection of informational dynamics in crowdsourced big crisis data produced by social media users is one of several avenues for socio-technical innovation that seeks to bridge between connected communities in crises and professional responders (see also ). In a recent study on the relationship between ‘real volunteer groups’ who respond to a crisis locally and ‘virtual volunteer groups’, who mainly carry out their activities online , Reuter et al. argue that, since direct collaboration between connected communities and official responders is difficult, it is important to enable digital volunteers to facilitate such collaboration. Efforts in formalizing the role of Virtual Operations Support teams (VOST) are already driving this agenda . However, to be able to support liaison between the different agencies in a way that recognises and respects connected communities as another, ephemeral, intensely connected and loosely self-organised agency, support for ‘bridging’ between different actors, systems and platforms is needed. Resilient ad-hoc networking support, systems of systems innovation and support for emergent interoperability is essential for this. With in the crisis management IT innovation community, there are numerous efforts to this aim. The BRIDGE project, which is the context for the research presented here, is one of them (http://www.bridgeproject.eu/en).