In the push towards bridging humanitarian efforts and advances in computing and artificial intelligence, there seem to be a minimal number of thinkers AND doers. Dr. Soenke Ziesche recognizes the imperative need to better integrate these worlds and sees potential implications for both those in the humanitarian fields and those in the AI sectors.
As a member of the United Nations since 2000 in the humanitarian and recovery sector, Dr. Ziesche speaks from a grounded perspective. When asked to speak about the overlap between these two fields, he opts to take another perspective, that there exists quite a gap between humanitarian issues and the field of AI. Applications of AI have mostly been limited to western countries. While there may be more isolated attempts to try and apply technologies for humanitarian-based reasons, a tangible bridge that unites the two has yet to be implemented. Perhaps the humanitarian field has been too conservative; or, perhaps the AI field has had too narrow of a scope in its outreach efforts. Likely it’s some of both of these issues, and the fact that there are not many people who are actively involved in both fields; often when there’s a gap, there’s a lack of communication between two groups, and this seems to be the case between humanitarian activists and AI scientists.
Granted, there are many components of humanitarianism, as with any field. The UN’s homepage for Humanitarian Affairs includes a list of thematic issues – everything from demining to global food security to protection of civilians in armed conflict. While any of these areas could potentially be helped by advanced computing and developing artificial technologies, Early Warning and Disaster Risk Reduction is an area that often demands greater attention. Unfortunately, even though some gains have been made in this area, there remains a need for the leveraging of technologies, particularly in the areas of communication and coordination between managers of crises’, aid workers on the ground, and victims of a disaster.
Mobile phones are invaluable tools in relaying information from disaster areas. Click the link on the UN’s page for the Global Disaster Alert and Coordination System (GDACS) and find a homepage that shows a real-time list of current emergencies and alerts from around the world. A page for Mobile Technology for field operations states that smartphones are becoming widely available, even in remote countries, and have the ability to provide important information from disaster areas; however, the implementation and management of such technology presents manifold obstacles. Beginning in 2011, a conglomerate of organizations began working towards solving some of these issues, including development of user-friendly tools for different populations; a common application programming interface (API); a system for culling and processing an array of information; and a plan for how and when to promote tools in the GDACS community.
Dr. Patrick Meier is a lead thinker and navigator in the area of applying technologies for early warning crisis and humanitarian response and resilience. Presently Director of Social Innovation at the Qatar Computing Research Institute (QCRI), Dr. Meier has written extensively about and is currently working towards a research-based framework for an information system for crisis response and management that he dubs Next Generation Humanitarian Technology & Innovations. There do exist humanitarian donors and organizations with investments in technology – DfID, ALNAP, and OCHA are a few that he mentions, but he also pinpoints the crucial missing link as familiarity by many humanitarian agencies with field of AI. Meier also clearly articulates that mobile phone use has sky-rocketed, and social media sites such as Twitter are heavily used amidst conflicts and disasters. But as attributed in a quote by DfID, “…Currently the people directly affected by crises do not routinely have a voice, which makes it difficult for their needs to be effectively addressed.” What’s more, he also affirms the point that in the face of disaster, mass amounts of data pour in quickly, and analysis of this data – like food – has a “sell-by” date. Having systems by which to organize and make sense of massive amounts of data is critical.
Dr. Meier believes we have the tools to effectively begin to address the “Big Crisis Data Challenge”, and that they are not unique or new ones; we need to make better use of Human Computing (which he sub-defines as crowd-sourcing and micro-tasking) and AI (sub-defined as natural language processing and machine learning) in mitigating these challenges. His far-reaching philosophy is that relevant technology applications within both of these methodologies must be united by a framework that promotes Research and Development (R&D) and is applied to humanitarian response and crisis prevention.
It seems that exasperating the communication gap between fields is the idea that disasters often serve as the triggers toward action; discussion and speculation about which disasters could or might occur often exhibit a theoretical or more visceral tone of reaction. In the immediate wake of publicized disasters or conflicts, people’s interests or emotions are often peaked, and actions toward prevention of the same types of disasters or conflicts are frequently undertaken with renewed vigor by a greater segment of the governments, other organizations, and the scientific community. As voiced in this article published by Ovum, a data analysis company, connection technologies that aid in a response to humanitarian disasters has become an established field, but using technologies in the prevention of such disasters is much more challenging.
Another important point is that AI could potentially be the cause of catastrophes. A well-developed humanitarian structure could be used to address these potential risks, but populations will not be prepared if the AI field fails to give as much attention to the inherent risks of AI as they do to the potentials. How do we plan for catastrophes at levels we haven’t seen before? Dr. Meier and other experts in the humanitarian technology field are honing in on how big data can be leveraged in both structural and operational crisis prevention. The Ovum article gives a succinct overview of how such data can be used in both types of prevention, taking into account data from three distinct phases in the “life” of a crisis – pre-event, during, and post-event. As Dr. Ziesche mentions, the field of AI has as much to learn from the humanitarian sector as the other way around. Understanding these three phases of data in specific geographic, cultural, and demographic contexts, and how they evolve, will provide an important window through which to consider and prepare for the prevention and intervention of natural, social, and AI-related disasters.