There are plenty of stories about artificial intelligence ending the world as we know it (see 2001: A Space Odyssey, The Terminator, certain Elon Musk tweets). There are fewer about how AI could save the world, but believers are out there. At a gathering Monday for the two-year anniversary of the Paris climate accord, Microsoft announced a $50 million investment in its AI for Earth project that it believes can be a “game-changer for our planet.”

Microsoft launched AI for Earth over the summer with an initial $2 million, hiring Lucas Joppa to run it as the company’s chief environmental scientist—a role that, as far as Joppa knows, doesn’t exist at any other technology company. Since the launch, the company has given out 35 grants to groups in 10 countries. The new investment will support and expand upon the current projects as well as new ventures over the next five years.

“We’re impatient and a lot of partners are too,” Joppa told MIT Technology Review. “There isn’t enough money and resources for people doubling down on these issues.”

AI for Earth grants give groups access and training to use the Azure platform and other Microsoft AI products. The projects so far include producing a more precise land cover map of the Chesapeake Bay watershed for conservation work, a mosquito tracking program that lets researchers learn about wildlife populations through blood analysis (it also functions as an early-warning system for Zika outbreaks), and getting farmers more data to increase crop yields with fewer resources.

AI for Earth is, of course, good for Microsoft’s bottom line as well. Joppa said that if the technology works for small, cash-strapped groups often working in tough conditions in the field, Microsoft gets feedback for their products and clients can be assured that the system will work for them.

“Anytime we can stress-test the technology, the better,” he said.

And of course, AI talent retention and recruiting is a huge concern for the tech industry. Joppa said that AI for Earth and its mission are a draw for researchers and developers.

“Yes, you need to pay them,” he said. “But you need to give them meaningful things to work on.”