Portal de Periódicos da CAPES

Human-computer interaction and social media in safety-critical systems-Editorial

2017; Wiley; Volume: 26; Issue: 1 Linguagem: Inglês

10.1111/1468-5973.12192

1468-5973

Christian Reuter, Tilo Mentler,

Occupational Health and Safety Research

Human–computer interaction in safety-critical systems in general, and the application of social media in safety-critical events, such as disasters and emergencies, are an interdisciplinary challenge. However, what are safety-critical systems and what are challenges for safe and efficient human–computer interaction? Safety-critical systems are systems that are relevant for safety and whose failure result in huge consequences. As humans are in charge of controlling and monitoring complex human-machine systems (Herczeg, 2000) and as the number of system failures resulting from design errors is underestimated, human factors issues have to be investigated not only, but also especially in the context of safety-critical systems (Westerman & Hockey, 1997). In doing so, the early work “Human Factors in Safety-Critical Systems: An underestimated contribution” of Westerman and Hockey (1997) emphasizes that application methods have to be supplemented by the use of human factors techniques to reduce errors throughout the design life cycle. Usability engineering integrates methods of software engineering, cognitive engineering, and software ergonomics to create usable, that is effective, efficient and satisfactory, interactive systems by analysing contexts of use, collecting requirements of users for the conception, and evaluating early and advanced prototypes in an iterative manner. Design-for-error (Herczeg, 2000), as an example, is necessary in the fields of safety-critical systems because designers and developers have to take extraordinary circumstances of system usage into account. To address the design-for-error, a selection of human factors principles is given by Leathley (1995) towards learning, alarms, and trade-offs for the menu design, consistency versus reduced keystrokes, and modality. Usability and user experience have to be considered as well (Mentler & Herczeg, 2016). However, while talking about mobile and ubiquitous devices used, for example by citizens (Helsloot & Ruitenberg, 2004), infrastructure breakdowns might also appear, and resilient approaches are needed to address technical issues (Reuter, Ludwig, Kaufhold, & Hupertz, 2017) as well as the perception of the use of these media has be studied to be able to address them in an appropriate manner (Reuter & Spielhofer, 2016). These are all principles that should be applied in safety-critical systems of different kinds. Situation assessment, social media analysis, sense-making, classification of information, crisis communication, and crisis informatics are examples—all dealing with the use of social media (Palen & Anderson, 2016)—that will be addressed in this particular issue with the specific focus on how to design human–computer interaction in safety-critical systems. We have thankfully received a significant number of submissions, and after peer reviews, four contributions from researchers from different countries (Australia, Germany, Sweden, Spain, and the USA) across different continents (America, Australia, Europe) have been accepted. The article “Sense-Making in Social Media during Extreme Events” by Stefan Stieglitz, Deborah Bunker, Milad Mirbabaie, and Christian Ehnis (University of Duisburg-Essen and The University of Sydney) deals with the use of social media for making sense of an extreme event. Besides possible social media usages as communicating, self-organizing, managing, and mitigating risks, the authors focus on the analysis of commentary-based social media communication practices during the Sydney Lindt Café Siege in 2014, the Germanwings plane crash in 2015, and the Brussels Terror Attacks in 2016. In doing so, they provide empirical data for better understanding intersubjective sense-making by identifying sharp differences in commentary-related communication structures on Twitter during these events. For investigating crises communication, they combined statistical, sentiment, and social network analysis methods and revealed that certain active user groups had an impact on the intersubjective sense-making. The second article “Disaster Response Aided by Tweet Classification with a Domain Adaptation Approach” by Hongmin Li, Doina Caragea, Nic Herndon and Cornelia Caragea (Kansas State University and the University of North Texas) deals with the classification and extraction of valuable information from social media platforms such as Twitter for improving situation awareness and aiding disaster response during emergency events. As labelled data (tweets) from the event of interest are hardly available for supervized machine learning algorithms, unlabelled tweets are quickly accumulating. Although labelled tweets from a previous disaster (called source) are presumably available and can be used to build a supervized classifier for the current disaster (target), it will not be able to make use of features specific to the target disaster (e.g., words used only in relation to the target disaster). This limitation leads the authors to propose a domain adaption approach, which learns a classifier for a current target disaster based on labelled data from a previous source disaster as well as on unlabelled data from the target disaster itself. For this, Naïve Bayes learning algorithm is used as the base classifier and an iterative self-training strategy to transfer knowledge from source to target. By identifying tweets that are relevant to the disaster of interest, experimental results show that the domain adaptation approach results in better classifiers compared to the supervized classifiers. The third article “Informing Crisis Alerts using Social Media: Best Practices and Proof of Concept” by Joel Brynielsson, Magdalena Granåsen, Charlotte Hellgren, Sinna Lindquist, Susanna Nilsson, Jiri Trnka and Maribel Narganes Quijano (FOI Swedish Defence Research Agency and Tecnalia Research and Innovation) shows that online sharing and social networking have had an impact within the field of crisis communication, and the way crisis information is disseminated and updated. Their paper addresses issues related to the use of social media for communicating crisis information and broadcasting alerting messages to the general population. A systematic literature review was made to identify important factors that affect the use of social media for alerting. These factors were mirrored in crisis management organization experiences, collected through interviews. The factors (timing, timeliness of response, trust) formed the basis for suggestions regarding the design of technological tools in general, and the design of a sentiment analysis screening tool for crisis alerting in particular. The results are part of a long-term research effort and have been used for guiding the development of a proof of concept system. The fourth article “15 Years of Social Media in Emergencies: A Retrospective Study and Future Directions” by Christian Reuter and Marc-André Kaufhold (University of Siegen) aims to recapitulate the establishment of social media in many larger emergencies and crises since the terrorist attacks of 9/11. During the last decade, sometimes summarized under the term crisis informatics, a variety of studies focusing on the use of ICT and social media before, during or after nearly every crisis and emergency has arisen. Their article reviews 15 years of social media in situations of emergency and its research with a special emphasis on use patterns, role patterns, and perception patterns that can be found across different cases to point out what has been achieved so far, and what future potentials exist. Human–computer interaction and especially applications of social media play a major role in contingencies and crisis management. This particular issue tried to highlight achievements as well as current trends. We are confident that this area of research will be of importance also in the future.