In this Sept. 21, 2018, photo, men deliver U.N. World Food Programme aid in Aslam, Hajjah, Yemen. The United Nations and independent donors are rushing food to a desperate corner of northern Yemen where starving villagers were found to be living off leaves. (Hammadi Issa/AP)

In September, the World Bank announced a new partnership with the United Nations, the Red Cross, Microsoft, Google, and Amazon to help end famine. The group will use artificial intelligence and machine learning to predict famine and mobilize early funding to mitigate its effects. The goal of this Famine Action Mechanism (FAM): Ready sufficient funding to respond promptly to food crises.

FAM is part of a new trend of donor-funded programs that try to avert emergencies by responding before crisis hits. But donors and aid organizations already have early warning systems in place. Humanitarian responses to food crises are chronically underfunded. Organizations delivering aid frequently ask donors to give more so they don’t have to reduce the amount of food they provide to the suffering. Warnings that are even earlier aren’t going to help boost funding.

So what will? My research suggests the biggest barriers to preventing famine are not technological, but governmental.

What problems can technical solutions solve when it comes to famine?

There’s a saying that when you have a hammer, everything looks like a nail. Technology firms can solve problems caused by a lack of timely, accurate data. And so that’s what FAM does. Its proponents argue that collecting all available data using remote sensing technology and analyzing it with machine learning and artificial intelligence tools will more precisely predict when and where food crises or famines will occur. Advocates say this will allow them to mobilize resources and respond more quickly, reducing both the severity of the emergency and the number of people affected.

But existing international initiatives already analyze conditions in famine-prone countries and predict onset. For instance, the Famine Early Warning System Network (FEWS NET), a USAID-funded initiative, forecasts food security conditions up to 12 months in advance. Systems like FEWS NET measure social and environmental factors — weather, market prices, migration patterns, harvest yields — to determine vulnerability and predict the likelihood of future emergencies.

International donors also fund governments to collect data and analyze their own conditions. Governments use these food forecasts to predict whether or not they’ll have a shortfall the following year. They can then assess budget priorities and request assistance, as needed.

And that’s key. In the end, the decision of whether, how, and when to respond are political choices made by governments.

The World Bank and its partners acknowledge these existing initiatives, saying they will build upon these approaches. But will FAM change how governments behave — whether or not they ask for help in time to prevent disasters — and how donors respond?

Governments and donors both politicize famine information

Unfortunately, FAM could complicate efforts to prevent famine. That’s because having several sets of famine-warning data and analyses means that governments, funders, and aid organizations can pick the data and analysis that support their preferred plan of action.

For instance, sometimes governments accuse humanitarian groups of exaggerating a country’s needs just so those groups can raise funds and prove their own importance — and implicitly, government’s weakness. Similarly, aid groups claim that governments’ analyses underestimate risks to protect those governments’ reputations.

No matter how impartial or rigorously collected, data and analyses can be interpreted in politicized ways.

How does aid get to the people in need?

My research investigates the relationship between humanitarian NGOs and host-country governments, examining how, when, and why governments and officials encourage or block aid delivery. Toward that end, I’ve interviewed more than 30 funders, policymakers, and others working on emergency response operations in Washington, D.C. and Niger.

I learned that, first, to deliver aid, humanitarian groups have to come to some kind of agreement with national governments. All aid organizations must have at least tacit government permission; U.N. agencies require formal permission. Aid agencies need to secure visas for their staff, register with the government, and establish the right to import supplies.

Governments do not always accept this support. They can limit any part of the aid that they wish — who delivers it, where it gets delivered, and what kind of aid is delivered — both through bureaucratic barriers and through violence. And in fact, recent research finds that more and more governments are limiting any aid delivered by foreign-funded NGOs.

Governments can revoke permission at any time, for any reason. In some cases, governments deny that there’s a food crisis or seriously underestimate its severity to preserve the government’s reputation and help keep its hold on power. Alex de Waal has chronicled cases in which governments cause or worsen famines.

Sometimes the system does work 

Of course, sometimes the system works well. Governments heed the early warnings, and ask for aid well before the expected food crises. Prompt donor support enables them to have food supplies at the ready in areas likely to need them; to offer short-term jobs that pay cash; and to sell food at a reduced price. This reinforces the market as the primary source of food while still assisting those in need.

But why don’t all governments do this? Amartya Sen’s famous work, recently upheld with empirical support, suggests that two things are needed to force governments to act to prevent famine. Accurate and useful information is one; the other is ways to hold governments accountable for their actions — through protest, media, and elections. Failing to prevent famine has at times toppled governments. Governments accountable to their people are more likely to invest in anticipating emergencies and requesting aid.

But donors must also respond with prompt, sufficient funding. This rarely happens. The World Food Program estimates 74 million people currently require need emergency food assistance — and donors have not given enough to feed them all.

FAM may improve the data, but it can’t convince recalcitrant governments to prevent famine, or push donors to give more. Given limited resources, creating yet another data collection and analysis tool may not be the most effective investment; improving existing systems would likely be faster and less expensive.

The bigger question to consider is this: How will the world offer incentives that prompt governments to deliver the needed assistance when their people suffer?

Allison Namias Grossman (@allisonnamias) is a PhD candidate in political science at the University of California in Berkeley, where she is completing a dissertation on foreign assistance and domestic politics, with a focus on West Africa and the Sahel.