The American Red Cross: Could the Future of Disaster Response Involve Drones and Machine Learning?
The American Red Cross is leveraging technology to improve the accuracy and timing of its critical information flows in disaster response efforts.
On August 23, 2005, Hurricane Katrina caused catastrophic damage in Louisiana and the Southern United States. Katrina’s winds and twenty-seven-foot storm surge caused $81 Bn of damage and over 1,800 lives lost [1]. Amidst the heartbreaking devastation, emergency response teams were slow to reach the citizens in need. Responses by both government agencies and non-profits such as the American Red Cross were widely viewed as insufficient in the wake of such a massive disaster.
Today, organizations like the American Red Cross are eager to leverage technology to improve response times to large scale natural disasters and empower local communities with easily accessible preparedness information. The hope is that if faced with another Hurricane Katrina, quicker response times could save even more lives.
The American Red Cross
The American Red Cross responds to over 70,000 disasters in the United States each year, as well as working internationally with global networks to deliver lifesaving services [2]. The Red Cross works in partnership with government agencies to provide services to disaster victims including shelter, food, health, and mental health services. As part of its mission, the Red Cross promises to “always be there in times of need” [3], and its business model depends on being among the first responders to a natural disaster and mobilizing its people and donor capital towards that effort. To quickly and accurately respond to global natural disasters, the Red Cross relies on maintaining accurate real time information flows that help the organization make decisions on where and how to deploy people and physical supplies so that they can truly always be with those in need.
Technology is Solving Traditional Information Flow Barriers
A large problem the Red Cross faces in regards to information flows is obtaining accurate information on the location of communities within affected disaster zones. Hundreds of millions of people are currently unmapped by traditional census tools, particularly in informal low-income communities in major urban areas [4]. To incorporate these citizens into disaster response planning, the Red Cross started the Missing Maps project in conjunction with several other humanitarian organizations. For two years, 16,000 volunteers used satellite imagery to identify buildings, infrastructure, and geographic features that helped put 20 million people on the map [4]. This past year, Facebook sped up this progress by sharing its population density data with the Missing Maps project. Facebook incorporates a machine learning layer that is purely trained on binary data of whether an image contains a building [5]. The model has proven to be very accurate in sparsely populated areas such as Malawi, in which Facebook predicted houses with 94% accuracy [5]. Together, Facebook and the Red Cross hope to map an additional 200 million people [4].
Aside from knowledge into distribution of local communities, the Red Cross also needs to have real time information as to where the effect of the disaster is being felt. Using grants from Dell, the Red Cross created three Digital Operations Centers that leverage social media to respond to natural disasters [6]. The centers use Salesforce’s Marketing Cloud to listen to social media conversations happening in disaster-affected areas and incorporate that on-site data into their disaster response decision making, including where to position people and supplies [7]. Social media content such as affected street addresses, on the ground photographs, and real time road conditions can save response teams valuable time and increase the overall effectiveness of the Red Cross. In addition, the Red Cross can share this information with government agencies to increase effectiveness of the entire disaster response effort.
The Next Frontier of Technology in Disaster Response
The Red Cross has effectively used partnerships with Facebook, Dell, and Salesforce to leverage technological innovation in obtaining real time data that allows for faster and more accurate disaster response. Interesting possibilities for the Red Cross to consider going forward include:
- Drones: The Red Cross recently commissioned a new study on drones in disaster relief efforts by Measure, a global provider of drone services [8]. While the Red Cross is understandably cautious regarding the regulatory uncertainty that still surrounds commercial drone use, the study details several interesting benefits drones could bring to the table. These benefits include furthering access to real time information at disaster sites through drone video capturing capabilities and delivering light-weight physical goods to hard to reach areas.
- Machine Learning: While the Missing Maps project already incorporates a degree of machine learning in Facebook’s mapping capabilities, The Red Cross could further incorporate machine learning into its operating model. One example of an area that would benefit from machine learning is assigning weights to the social media information gathered by the digital operations centers, so that Red Cross personnel is given the most relevant information to each disaster response task, instead of having to comb through a vast quantity of digital information.
(799 Words)
Sources
1 “Hurricane Katrina Led to Largest Ever Red Cross Relief Response,” August 28, 2015, American Red Cross, http://www.redcross.org/news/article/Hurricane-Katrina-Led-to-Largest-Red-Cross-Relief-Response, accessed November 18, 2016.
2 American Red Cross, “What We Do,” http://www.redcross.org/what-we-do, accessed November 18, 2016.
3 American Red Cross, “Mission & Values,” http://www.redcross.org/about-us/who-we-are/mission-and-values, accessed November 18, 2016.
4 “Data from Facebook Helps Red Cross Make Better Maps,” November 15, 2016, American Red Cross, http://www.redcross.org/news/article/Data-from-Facebook-helps-Red-Cross-make-better-maps, accessed November 18, 2016.
5 Tobias Tiecke, “Open Population Datasets and Open Challenges,” Code (blog), Facebook, November 15, 2016, https://code.facebook.com/posts/596471193873876, accessed November 18, 2016.
6 “How Ten Years Changed Disaster Preparedness and Response,” August 19, 2015, American Red Cross, http://www.redcross.org/news/article/How-Ten-Years-Changed-Disaster-Preparedness-and-Response, accessed November 18, 2016.
7 Salesforce, “American Red Cross,” http://www.salesforce.com/customers/stories/redcross.jsp, accessed November 18, 2016.
8 Matt McFarland, “American Red Cross Takes Serious Look at Using Drones for Disaster Relief, Holds Off for Now,” April 21, 2015, The Washington Post, https://www.washingtonpost.com/news/innovations/wp/2015/04/21/american-red-cross-takes-serious-look-at-using-drones-for-disaster-relief-holds-off-for-now/, accessed November 18, 2016.
I think the benefits of machine learning and drone technology for disaster relief are enormous, and I’m so glad you shared this, Ana! However, I think there are still quite a few challenges that we need to overcome before we can really consider this technology as part of the solution for disaster relief. First and foremost, commercial drone operations are currently illegal in the United States without an FAA-granted exemption, and rules likely won’t be finalized until 2017. More importantly, because the Red Cross depends so heavily on donations, it will be critical for the organization to tell a cogent story that convinces donors of its plans to digitize disaster relief. I was surprised to find that only 21% of Americans favored commercial use of drones in a 2014 AP poll [1]. Finally, a third concern I have is just regarding the training and managing of drone systems. In order to make this work, the Red Cross not only needs to learn how to implement drones into their systems, but also must essentially create an entirely new airline industry of in-air drone management. With so many other industries and companies also looking at drone technology, I would imagine this might get complicated.
[1] https://www.washingtonpost.com/news/innovations/wp/2015/04/21/american-red-cross-takes-serious-look-at-using-drones-for-disaster-relief-holds-off-for-now/
I think you bring up a good point here, Catherine– we have a long way to go for drones to be socially acceptable or under regulatory control– but I think this also presents an opportunity for the American Red Cross.
As drone prices have dropped, the question now is “what can we use them for?” The answers are almost unlimited, but there has yet to be drone integration into everyday business operations that is truly groundbreaking. If the American Red Cross takes the results of the Measure study and uses the existing FAA 333 Exemption and Part 107 Rule to build out initial uses for drones in its operations, it is poised to throttle these operations into full gear when more complete FAA regulations are released. Commercial drone use in disaster aid could become a crowded space quickly but the American Red Cross can claim first mover advantage if they take the right investment steps now.
More broadly, I think this post highlights the challenge of regulation on digitization– could the American Red Cross already be saving more lives and providing more aid with drones if the FAA wasn’t standing in the way? Yes, the FAA’s regulations will likely provide some protections to Americans from unscrupulous drone use, but the FAA’s current inability to pass these regulations is stifling the market drone technology creates. Digitization is a huge positive for society, but it can only create positive outcomes if it is not unduly hampered by government intervention.
Ana – It’s so cool to learn about the partnerships with the ARC has established with leading technology players, such as Facebook. One challenge for ARC that isn’t captured in this post is that its access to data often depends on cellular connectivity and / or internet access – two modes of communication that are often unavailable during times of crisis. A key question for me is how ARC can partner with organizations or invest in new technologies to enable communications among crisis-afflicted populations.
Additionally, it’s interesting to consider what the role of non-profits and for-profit tech firms are in this space: Do you think it’s appropriate for ARC to be directly investing in digital technologies or should it just be waiting to partner with best-in-class tech companies and benefit from their advancements?
Sources:
[1] https://www.wired.com/2016/11/facebook-disaster-response/