At NASA, AI is helping scientists unlock discoveries hidden in decades of data

The goal is to provide access to information that once required significant expertise - a farmer, for example, can ask Hydrology Copilot what the environmental conditions were that year in a particular area. Local agencies can explore water conditions in their areas without needing technical expertise.

"I truly think this is going to democratize (data)," Kumar says. "It goes beyond scientists and practitioners who know how to work with the data to the average person who can just access the data and the information that flows from it."

AI is also transforming weather forecasting, Kumar says. NASA foundational models trained on decades of Earth science data and physical models can help detect and simulate weather patterns quicker. AI is also enabling faster responses to disasters by allowing satellite data to be processed on the edge, close to where it's created or used, instead of needing to transmit it back to Earth first.

"Let's say we are looking at the Earth and there's a flood happening. By the time you beam the data down to Earth and we process it, it's probably too late," Kumar says. "If you're able to process the flood map on the edge and give you that information directly, that's a matter of minutes. And that makes a huge impact."

'The possibilities are just endless'

NASA's growing interest in AI was on display during its first agencywide data and AI hackathon last year. Steers-Smith was advised to expect around 30 participants. More than 500 employees signed up.

Among the winning projects was a system of AI agents that uses a chat interface to let users search NASA's extensive "lessons learned" public database, which holds knowledge from decades of missions and programs. Another used machine learning models to develop a solution for diagnosing the severity of forest fires to better protect firefighters and equipment.

Steers-Smith says the hackathon highlighted a growing interest among NASA employees in using AI in their daily work. "If you used AI to solve somebody else's problem, maybe you're better now at critically thinking about how AI works with the problem you have," she says. "I think it's an evolution of AI and where people are finding value in it."

For Steers-Smith, efforts like the hackathon underscore the value of AI to connect information spread across NASA's many data repositories. Future endeavors such as building the Moon Base, she points out, will require scientists and engineers to gather information from a large array of sources on factors ranging from seismic activity to solar conditions and ice deposits.

"We collect so much data. We have new data, old data, satellite data, images, video, audio recordings. And it's very, very hard to catalog all of that and connect it," she says.

"If we can use AI to start connecting those things, we'll make so much more discovery. The possibilities are just endless."

Header image: The Artemis II crew captured the Moon with a distant Earth setting in the background during the lunar flyby in April 2026. Image credit: NASA

Second and third image from top: A brilliant concentration of stars takes center stage in this image captured by NASA's James Webb Space Telescope. NASA's Perseverance rover took a self-portrait against a rocky outcrop on Mars' western frontier. Image credit: NASA

Fourth image: NASA's Space Launch System rocket launches from Kennedy Space Center in Florida, carrying the Orion spacecraft during the Artemis II mission in April 2026. Image credit: NASA

Story published on Sept 28, 2026

Deborah Bach writes about AI, innovation and the transformative ways organizations and people are using technology. A native of British Columbia, Deborah was previously a newspaper reporter for the Seattle Post-Intelligencer and the Baltimore Sun. Her work has been published in outlets including the New York Times, Vancouver Sun and TODAY.com, among others. You can reach Deborah on LinkedIn .