In the Battle Against Climate Change, Can AI Help Us Win?

What jumps to mind when you think about AI? Chatbots? Meeting assistants?

Yes, but instead, picture this. A team of scientists are in a room, squinting at grainy satellite images on their screens. They’re analyzing forest restoration, trying to find dots that would represent tree cover, but they’re hampered by an unreliable algorithm.

Exploring possible solutions, an idea sparks. It occurs to them that there’s another type of analysis that involves looking for tiny spots in grainy images: early-stage brain cancer detection.

So the scientists, who worked — and still do — at the World Resources Institute, followed their hunch. They adapted the AI used to detect cancer cells in brain MRIs and applied it to their own research — and guess what? They found trees covering an additional 180 million hectares of land than previous data had shown. That’s an area half the size of India.

I love that story because it illustrates the potential that AI has to help us solve multiple problems. If we can harness that potential, and apply it responsibly to climate and nature, imagine the progress we might make in protecting our planet.

At the Bezos Earth Fund, we are investing boldly in that potential. In April we announced our AI for Climate and Nature Grand Challenge, a commitment of up to $100 million to support the most promising ideas for leveraging AI to fight climate change and nature loss.

The first round of awards will focus on three specific areas: Biodiversity conservation, sustainable proteins, and power grid optimization. And a “wild card” category to bring to life powerful ideas that may not fit perfectly within any of the three areas.

In each of these, we’re hoping to discover answers to practical problems around the world.

For example, can AI help conservationists protect and restore nature? Scientists could use AI-enabled camera traps and bioacoustics tools to monitor endangered species — which are often elusive and difficult to track — and rapidly take action to preserve them. Field practitioners may find many other opportunities to use AI, from designing wildlife corridors and protected areas to identifying invasive species.

Or can utility providers across the globe us AI to mitigate power outages and increase affordable access to electricity? AI can help power companies quickly analyze vast amounts of data from smart meters, weather forecasts, and consumer behavior to predict energy demand and inform preventative grid maintenance. AI solutions may help us rapidly transform our grid to optimize two-way flows of electricity among millions of small generators and batteries millions to lower costs, expand access and become less reliant on fossil fuels. In the process, we could make it more resilient and inclusive, so that no person gets left behind in a clean energy future.

And what about feeding the world? The human population is expanding, yet agriculture is one of the biggest strains on our planet. Can AI help sort through the millions of possible protein combinations to find the ideal ones to produce nutritious meat alternatives — just as tasty as beef, but cheaper and much better for the environment?

These and hundreds of other questions reflect the seriousness of the problems we are facing and the urgency of our collective need for solutions. They also reflect the sense of possibility that fueled us to launch the AI for Climate and Nature Grand Challenge. We are living in profoundly challenging and exciting times. We don’t know how much AI can contribute — but in this decisive decade, it is critical that we find out.

And yes, we need to be vigilant in managing the possible abuses of AI and its potential negative consequences. But we must not allow these risks to prevent us from exploring how we might use it to protect this wonderful planet.