Jyunmi Hatcher

That readiness is partly the crew and partly the layer of automation underneath them.

So the autonomy on Orion is sophisticated, but it's deliberately conservative.

Radiation hardened processors, several generations behind consumer chips, classical control algorithms, and no large language models running anywhere on the spacecraft.

NASA's other ongoing experiment in space autonomy is more ambitious, and it's already operational on another planet.

So in December of 2025, NASA's Perseverance rover completed the first drives on another world that were planned entirely by generative AI.

JPL announced that the demonstration in late January.

On December 8th and December 10th, Perseverance drove a combined 1,496 feet across the Jezero crater on Mars using waypoints generated by vision language models, a type of generative AI that can analyze images and produce structured outputs.

The work was a collaboration between JPL's Rover Operations Center and Anthropic using the company's CLAWD AI models.

The AI analyzed the same high-resolution orbital images and terrain data that human rover planners normally use.

They identified hazards like boulder fields, sand ripples, and generated continuous driving paths.

Before the commands were sent to Mars, the engineering team ran through them.

JPL's digital twin of Perseverance, virtual replica of the rover, and verified more than 500,000 telemetry variables to make sure the AI's plan was compatible with the rover's flight software.

For the past 28 years, across multiple Mars missions, rover routes had been planned manually by human drivers.

The December demonstration was the first time that work had been done by AI on a mission already on the surface of another planet.

Vandy Verma, the space roboticist at JPL and a member of the Perseverance engineering team, said in the JPL press release, the fundamental elements of generative AI are showing a lot of promise in streamlining the pillars of autonomous navigation for off-planet driving perception, seeing the rocks and ripples, localization, knowing where they are, and planning and control, deciding and executing the safest path.

We are moving towards a day where generative AI and other smart tools will help our surface rovers handle kilometer-scale drives while minimizing operator workload and flag interesting surface features for our science team by scouring huge volumes of rover images.

Matt Wallace, manager of JPL's Exploration Systems Office, or ESO, connected the experiment directly to Artemis.

Imagine intelligent systems not only on the ground at Earth, but also in edge applications in our rovers, helicopters, drones, and other surface elements trained with the collective wisdom of our NASA engineers, scientists, and astronauts.

That is the game-changing technology we need to establish the infrastructure and systems required for a permanent human presence on the moon and take the U.S.

to Mars and beyond.