On September 5, 2014, flash-monsoon rains devoured the village of India-, China-, and Pakistan- controlled Kashmir within minutes, burying 80% of the territory. Panicked citizens fled to home roofs, where many would await rescue for up to a week without food or water. The hashtag #KashmirFlood became a lifeline for many as victims tweeted their locations, needs, and of missing loved ones in hopes of a salvation response.
Others were less fortunate. For them, almost two weeks passed before Internet connection was restored.
Once restored, people reported their statuses to loved ones via social media.
And some pleaded for information on the missing via Twitter and Facebook, with hopes their outreach would be shared and loved ones located or rescued.
On the other side of the world, on August 14, 2014 in Napa County, California, victims made thousands of 911 calls when a magnitude 6.0 earthquake hit the area. Still, entire regions couldn’t access help as damaged power and communication lines rendered distress calls impossible. The problem: people were served on a first-come, first-served (FCFS) basis. Victims with access to the necessary power and communication lines to call 911 were served first, leaving those with fewer survival resources to fend for themselves.
AI and ML Power Life-Saving Natural-Disaster Management
Meanwhile, Stanford University Engineering student, Ahmad Wani, shares his experience:
During a break from graduate school…I was visiting my parents in Kashmir when a large flood engulfed the state. 80% of Kashmir went under within minutes. Most people were on their rooftops for up to seven days without food and water…Most of the rescue was random, and response priorities were ad hoc.
But that’s not all. After surviving the devastation of his third-world-based home, he returned to his first-world alma mater to find an ironically similar scene: in the hard-hit — albeit first-world — Bay Area, an FCFS disaster-management approach meant aid efforts were also performed at random, with no priority given to those most at risk or in need.
Before, During, and After, ML and AI Turn Disaster Victims into Survivors
Recognizing the universal need for better disaster management systems, Wani’s vision for One Concern was born. A machine-learning (ML) and artificial-intelligence (AI) powered natural-response solution, One Concern gives responders tools and insights to take life-saving and efficient steps for full restoration prior to, during, and following a disaster event. Near real-time and even predictive reports of damage and impacted regions helped first-responders know what needs to prioritize and where to first allocate life-saving resources, information, and personnel.
One Concern isn’t alone in leaning on ML and AI to walk alongside disaster victims. Other tech companies like IBM Watson and WorkFusion, as well as government organizations are leaning on ML and AI solutions to turn victims into survivors at every stage of disaster management.
ML and AI Provide Preventative Insights, Making Way for Preventative Measures
Even before natural disasters hit, ML and AI solutions like One Concern’s power predictive simulation models, thereby training and otherwise preparing region-specific response teams to mitigate disasters’ end-to-end effects.
For example, the U.S. Center for Disease Control (CDC) reports that 50,000 people worldwide die yearly from the flu and flu-related complications, a phenomenon especially affecting third-world nations. To reduce death tolls and to inform citizens on preventative measures, China’s Shenzhen CDC aggregates real-time data on air quality, geography layout, economic conditions, population density, and population mobility. They then automate insight-production via ML an AI to predict which regions are most at risk and target them with location-based resources to prepare for and control disaster impact.
ML and AI Hinder Disaster Spread as It’s Happening
During a disaster, ML and AI, when combined, can help responders:
– track and predict disaster trajectory;
– identify location-based urgent needs and most impacted regions;
– gain predictive insight into what resources need to be allocated first and where;
– gather and uncover trends in user-generated social-media reports of needs, impact, death tolls, and missing persons;
– send life-saving aid to most impacted or at risk regions;
– send targeted danger warnings to potential victims;
– send location-based and timely safety information; and
– locate victims to coordinate their rescue.
Going back to our Napa-County-earthquake scene, powered by IBM Watson’s more recent partnership with Association-of-Public-Safety-Communications Officials, responders could have nixed the FCFS policy, replacing it with a most-urgent-need, first-response policy and, thereby save lives.
With this partnership, IBM Watson uses ML to analyze how most-urgent calls unfold, then pass this insight on to dispatchers and dispatch trainers. The end result: dispatchers can recognize most-urgent calls, prioritize them, and gain insight on caller location and likely need. In turn, they can provide timely safety information, resources, and response personnel to save more lives.
ML and AI Expedite Recovery to Mitigate Trauma
ML and AI tech brands are also partnering with insurance and fundraising firms to expedite recovery and, thereby, shorten aftermath suffering and trauma.
Insurance company Metlife Japan, for example, works to serve communities that are home to one of Reuter’s 14-top-deadliest earthquakes and one of USAToday’s top-10-deadliest tsunamis of the past decade. To best serve an at-risk population, MetLife partnered with WorkFusion.
WorkFusion brings AI and ML-powered products to enable unmanned database mining, auto-validation of claims policies, claim-to-policy fact matching, automated decision-making, and data transition to systems of record (SOR) for downstream payment. This means an 80-percent reduction in insurance-claim-processing time. Translation: reduced suffering time for victims.
Using ML and AI-powered WorkFusion products, RevUp, located in wildfire and earthquake-prone Silicon Valley, matches likely donors to non-profit causes at record rates and speeds. Then, their combined solutions inspire donors to create look-alike donors via social-media sharing, resulting in a ripple effect of automated, targeted, and profitable efficiency. Lastly, they automate donation tracking to ensure financial goals and, therefore, victim needs are met.
ML and AI Help Humans Better Help Humans
At every stage of disaster-management, ML and AI solutions offer disaster responders the necessary real-time insights and automated decision-making to replace FCFS disaster-relief management with a more efficient and scaled most-impacted, first-served approach.
As a disaster relief organization, you may not be able to follow in the footsteps of large government entities like China’s Shenzhen CDC to power in-house ML and AI capabilities. But brands like RevUp and Metlife Japan offer a path forward: partnerships. WorkFusion, IBM Watson, and countless emerging AI and ML firms are eager to help you more effectively save and rebuild lives at every stage of the disaster-management journey.
Your homework: simply reach out to them. Work together to develop business cases that, when powered by ML and AI, can reduce natural-disaster death tolls and trauma impact.
This post originally appeared in TowardsDataScience.com.
