Six Ways AI Can Revolutionize Humanitarian Affairs

A change is coming

WFP/Gemma Snowdon Aid delivered in South Sudan

By Aditya Tiwathia

Aditya Tiwathia is a leader in humanitarian project development who serves as an advisor to the AI Policy Institute on issues of security and humanitarian policy, limits on the use of artificial intelligence (AI) in warfare, and the potential application of AI to peacebuilding.

The AI revolution in humanitarian affairs is upon us.

AI is set to usher in a world in which we can predict and prepare months in advance for conflict, disasters, floods, epidemics, low crop yields, or droughts. With Silicon Valley and the Pentagon are already all-in on AI, it’s high time we humanitarians get on board.

As a UN humanitarian, I have worked in the wake of war in Gaza clearing unexploded bombs, helped deliver shelter and life-saving items in the camps of Darfur, and participated in peacekeeping missions in conflict-affected South Sudan and Somalia.

I have also seen our essential humanitarian institutions buckling under rising chaos and human catastrophe. Over 300 million people need life-saving assistance in 2024, 86 percent of whom are facing acute food insecurity. Another 700 million are living on less than $2.15 a day. The humanitarian catastrophe of the Israeli-Palestinian conflict alone has added an additional 2.7 million people to the tally. That’s 1-in-7 people in the world living on the edge of precarity. 

Yet, nearly 60% of global humanitarian needs go unmet every year with the funding gap ever widening each year - leaving organizations like the UN overstretched and underfunded while crises fester under hastily-applied bandaids. What’s worse, these crises will only intensify over the next decade - driven by climate change and the ensuing rise in conflict, migration, resource competition, drought and food scarcity.

The only feasible solution is a complete paradigm shift: from the current reactive model to a truly proactive model based on prevention and preparedness. This is where AI can be uniquely valuable.

Until recently, terms like “prevention and preparedness” were just UN buzzwords limited to two forward logistics bases in Italy and Uganda. This disconnect between aspirational jargon and actual outcome was rooted in data scarcity. To understand what reality looks like, let me tell you about my experience at the World Food Programme (WFP) in Sudan in 2011. 

I arrived in Khartoum with six months to prepare for the predicted fallout from the referendum on the secession of South Sudan. Fresh out of graduate school, I wanted a data-driven preparedness plan guided by historical patterns and current data. For 6 months, I lost sleep and clumps of hair searching for a reliable approach to planning, one based on patterns of rainfall, roads' passability, past conflict, known migration routes, current census data, and other field reports. It quickly became apparent that “data-driven preparedness” was nothing more than empiricist kabuki for donors. The data either did not exist or was over a decade old. And, none of it could support reliable extrapolation to the situation at hand. Without data, our “preparedness plan” ended largely based on the self-interested and often unreliable testimony of local partners who (unsurprisingly) always recommended that humanitarian supplies were “best pre-positioned” right next to where they worked.

In the end, my worst fears came true. When the crisis hit, we only met 30% of the immediate needs of fleeing civilians. In some areas, heavy rainfall washed away roads, stranding much-needed aid in warehouses. Even where existing aid nodes had been established on the basis of prior patterns of displacement, we missed the mark. People did not move as they had done in the past - instead moving away from existing aid nodes to avoid being vulnerable to ambush attacks. Conflicts evolve rapidly. An effective humanitarian response required real-time data about population movements. Without that essential data, we were left unable to effectively meet the needs of those who depended on us.

That was thirteen years ago. Today, however, a data-driven humanitarian revolution is taking place. Most crises take place in data-rich environments, with nearly 80% of people, including the world’s poorest, most-desperate populations, emitting digital breadcrumbs through their mobile phones. Conflicts are broadcast on social media in real-time. And with AI, humanitarians now have the power to knit disparate data streams into actionable insights. 

Contrast my struggles in Sudan with UNHCR’s Project Jetson being piloted in Somalia. Using machine learning, Project Jetson can integrate market prices, rainfall patterns, and social media data to predict population movements due to famine or conflict months in advance. Similar projects, like the Danish Refugee Council's Foresight Model have used open data to precisely predict forced displacement in Afghanistan and Myanmar. Microsoft is collaborating with NGOs in India to predict cyclones using AI models. UNICEF’s project MERON is using facial recognition algorithms to detect early signs of childhood malnutrition. 

Such predictive systems are revolutionizing advanced financing, resource mobilization, preparation and potentially prevention. This is exactly what the International Federation of Red Cross and Red Crescent Societies (IFRC) is piloting via its Forecast-based Financing program. It is using meteorological and market data to preemptively predict the need for humanitarian resources. And, in 2020, advance prediction of monsoon flooding in Bangladesh led to anticipatory action and the fastest deployment of funding in UN history. 

To be sure, humanitarians must tread carefully. For humanitarians, errors on the way to innovation can have dire consequences. But unless humanitarians embrace AI, we have no shot at addressing the sheer magnitude of current and future crises.

We’ve seen what AI can do in these early days of its technological development. As we look the future, there are at least six ways AI can support a fundamental restructuring of the humanitarian sector and supercharge global humanitarianism.