Artificial intelligence could prevent famines, says World Bank president

The World Bank is developing the Artemis famine-prevention AI system with technology giants

The head of the World Bank believes that artificial intelligence could "end famine" by spotting potential famines before they develop.

Jim Yong Kim, who has led the Bank since 2012, told reporters that AI could provide as much as six months warning to aid workers before a crisis.

The World Bank is working with Microsoft, Amazon and Google on the famine-spotting system, known as Artemis. This tool can trawl through data - including weather reports, satellite information and even social media - and analyse it to draw conclusions.

"This could actually end famines," said Kim. "We are getting information well ahead of time instead of waiting until the fifth stage of famine...

"What we're talking about is going in prior to the first stage, getting pre-first-stage information... there are huge possibilities."

He was speaking at Stanford University in California, where students are helping to build the AI system. It is currently being tested in South Sudan, Niger, Mali, Chad and Somalia, and is planned to go live in "a small group of countries" by the middle of next year, says The Telegraph.

Instead of waiting for the UN to officially declare a famine (the fifth stage of food shortages), the AI could instead release relief money once certain thresholds are met.

Kim said: "One of the things that has been shown in Ethiopia is that if you think a famine is coming, if you do something really simple like double the amount of cash transfers that poor people get, you can actually stop the famine from going forward."

However, critics say that the AI could have difficulty in predicting political tensions and armed conflict, which are a significant contributor to modern famines.

"We have the best science in the world backing up the assumptions we make about rainfall," Chris Hillbruner, a senior adviser at hunger monitoring agency Fews Net, told The Telegraph. "It's a lot harder to build assumptions about what is going to happen with a conflict issue or a political issue six months or eight months in the future."