NASA is using AI to make decades of mission data easier to explore

NASA is using AI to help researchers find patterns in vast science datasets and to help mission teams make sense of operational information. The practical promise is less time hunting across scattered records, and more time putting them to work.

Watch Desk analysis

What happened

NASA staff use AI tools to search academic literature and support everyday work. Earth Copilot and Hydrology Copilot, developed with Microsoft, let people explore complex Earth science data through natural-language questions. Hydrology Copilot can analyse variables such as rainfall and soil moisture, with potential uses including understanding local water conditions and preparing for droughts or flood risks.

At Johnson Space Center, NASA has also developed a prototype AI agent for SPARTAN flight controllers, who monitor the International Space Station’s power and thermal-control systems. NASA describes the agent as a possible virtual member of the support team, not a replacement for the controller making decisions. The agency also says AI tools helped generate safety recommendations and risk-mitigation scenarios during Artemis II planning, work that traditionally took experts weeks.

The details come from a Microsoft Source feature on NASA’s use of AI, which includes comments from agency staff. Read the feature.

Why it matters

NASA’s challenge is not simply a shortage of data. Mission records, observations and scientific datasets are spread across sources and can be difficult to connect. Tools that help people search and analyse them could make existing information useful to more researchers, agencies and communities, without asking every user to become a specialist in data systems first.

The stakes change when AI moves from helping with research to supporting mission operations. NASA presents the SPARTAN system as a prototype under exploration; its account does not establish that it is making live flight-control decisions. That distinction is rather important when the subject is spacecraft systems, not a more convenient way to find a paper.

Our read

This is a more grounded case for AI than the usual promise that a chatbot will transform everything by Thursday. The value here is in joining up hard-to-navigate data and speeding up analysis. NASA’s examples are promising, but the operational agent is still a prototype, and the dramatic timing claims about Artemis II planning come from NASA staff quoted in Microsoft’s feature. Researchers and mission teams should judge these tools by the quality of their outputs and how people check them, not just by how quickly they produce an answer.

What to watch

• Whether NASA moves the SPARTAN agent beyond prototype exploration, and what role it is given.
• How NASA staff validate AI-generated safety recommendations and data analyses.
• Whether Earth Copilot and Hydrology Copilot make complex datasets more accessible in practice.
• What results NASA publishes on time saved and scientific discoveries enabled.

Discussion spark: Where should NASA draw the line between AI helping teams find and assess information, and AI influencing decisions in mission operations?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.