GenAI's role in a return trip to the moon and beyond

GenAI upgrades NASA's space program

NASA is all about developing and using the most advanced technologies to advance humanity on Earth and beyond. GenAI fits neatly in the agency's vision and provides significant benefits to the space program.

"GenAI is a qualitative step change in NASA's use of artificial intelligence, building on previous systems designed for specific tasks like navigation and data analysis," said Kevin Murphy, acting chief AI officer at NASA. GenAI models can create new content, synthesize information and engage in natural language communication. NASA is currently using GenAI to enhance science data analysis, support engineering through digital twins and improve mission operations by streamlining planning and documentation processes.

"This technology helps NASA increase efficiency and effectiveness across its core activities," Murphy said, adding that the agency uses GenAI with rigorous science and engineering guardrails while keeping humans in the loop. "These systems augment, they don't replace domain expertise," Murphy acknowledged. "Human judgment, validation and reproducibility are essential."

Back to the moon with a boost from GenAI

Humans last set foot on the moon in 1972 during the Apollo 17 mission. The Artemis program is NASA's program to return to the moon, and GenAI figures prominently.

The Artemis II four-person crewed, 10-day mission is a free return trajectory around the moon, comparable to what Apollo 8 accomplished in 1968. Artemis III will follow as an Earth-orbit test mission, comparable to Apollo 9's mission and is currently targeting mid-2027 as the launch date. The mission will involve rendezvous and docking in low Earth orbit with SpaceX Starship HLS and/or Blue Origin Blue Moon commercial lunar landers. Artemis IV is expected to accomplish a crewed lunar landing in early 2028, followed by the Artemis V moon landing in late 2028 and annual lunar landings thereafter.

In addition to launch vehicles, the Artemis program includes the planned Lunar Gateway, a space station to be assembled in orbit around the moon with the first module set for deployment as early as 2027. NASA's Gateway is intended to serve as a stepping stone to Mars and partner with space agencies from Europe, Canada, Japan and UAE.

Though the Artemis program has encountered delays and will likely continue to face challenges, GenAI will help accelerate the program in various ways. "Today, NASA's Artemis campaign is using GenAI to efficiently handle the tedious, time-consuming tasks that hinder programs," Murphy said. Those tasks include operations like combing through thousands of technical documents, requirements, simulations and telemetry streams, so experts can focus on the complexities of designing and planning the missions. "For Artemis," Murphy noted, "we are developing AI tools to aid in risk management and engineering compliance across the complex vehicles and operations associated with our missions."

Returning to the moon also involves future plans for a sustained human presence and colonization on the lunar surface. Murphy expects AI will play a key role in power systems optimization, life support, logistics and robotic operational partners.

"In the future, GenAI models can help with predictions about maintenance schedules, optimize resource distribution, prioritize scientific data to send to Earth and dynamically adjust schedules as the operational environment changes," Murphy explained. "Throughout the agency, NASA prioritizes rigorous scientific and engineering guardrails, ensuring that GenAI serves as a trusted, efficient partner that empowers human judgment, safety and resilience in demanding space environments."

GenAI finds a home on Mars

Since the NASA Perseverance rover landed on Mars in 2021, much of its operations have been painstakingly calculated by NASA engineers to navigate the alien terrain as part of the robotic system's mission. The rover's first drive on the red planet was planned by GenAI and executed on Dec. 8 and 10, 2025.

NASA's Jet Propulsion Laboratory partnered with Anthropic to use Claude AI vision-language models to analyze HiRISE orbital imagery and digital elevation models. The analysis identified terrain features, including bedrock, outcrops, boulder fields and sand ripples on the surface, to generate continuous paths with waypoints to safely guide Perseverance.

A one-way communication between Earth and Mars at maximum distance can take up to 24 minutes -- 48 minutes for a round-trip exchange. Real-time human control of Mars operations from Earth would not only be impractical, but also physically impossible.

NASA's Earth-Independent Operations Laboratory in the Human-Computer Interaction Group formally assessed ground support risks for humans during Mars missions as "code red" -- the most serious risk category. GenAI's ability to reason across multiple data types and generate novel responses to situations that were never preprogrammed is an operational necessity.

Lessons learned for businesses on Earth

NASA's use of GenAI for its space missions can potentially be applied to the way businesses use the technology on Earth. "NASA's experience deploying GenAI in the demanding environment of space provides a crucial blueprint for all high-stakes enterprise applications," Murphy said.

One key lesson is keeping humans in the loop. AI should be human-centered, Murphy suggested, to reflect "the right stuff" prevalent among the astronauts in the early days of the U.S. space program. "In low-margin-for-error operational environments such as aviation, space or medicine," he said, "it is humans who have the final say, and AI cannot be held responsible for safety."

Governance is another key lesson. In high-stakes environments, Murphy advised, AI must be deployed with strict governance and validation, focusing on trusted, efficient systems that empower human expertise and resilience. "Critically, we work within constraints, favoring efficient, trusted systems tailored to the task over merely implementing the largest model available," he explained. "[It's] a lesson in disciplined deployment that should guide every organization embracing this transformative technology."

Sean Michael Kerner is an IT consultant, technology enthusiast and tinkerer. He has pulled Token Ring, configured NetWare and been known to compile his own Linux kernel. He consults with industry and media organizations on technology issues.