The Perseverance Rover taking a selfie on a rock named Cheyava Falls. | Source: NASA/JPL-Caltech/MSSS
Location of the action: Mars, edge of Jezero crater, Sol 1707 (Martian day) of the mission.\
Event: The Perseverance rover made the first movements in the history of another world, the route for which was completely planned by artificial intelligence.*
Dispatcher: The Claude AI system from the company Anthropic, located on Earth, 360 million kilometers away.
🚗 How it was: the chronicle of the first taxi "order" on Mars
It all started when the mission team at NASA's Jet Propulsion Laboratory (JPL) decided to conduct an unusual experiment. Instead of manually studying the stereo images from the rover's navigation cameras pixel by pixel and plotting a safe path, they threw off the AI's task: "Claude, we need to drive about 400 meters through this rocky field near the crater. Choose the safest and most efficient route."
An AI trained to understand images and text (the so—called VLM - vision-language model) analyzed the panoramas of the area. He identified dangerous boulders, assessed the steepness of the slopes, and outlined a series of waypoints. This plan was uploaded aboard Perseverance, which then drove along it autonomously — first 210 meters, and then the rest.
Key detail: The AI did not control the rover in real time. He acted as a super-cartographer and tactician, creating a detailed plan that the rover then carried out on its own, using its usual autonomous driving systems to avoid unforeseen interference.
🎯 Why is this a breakthrough and not just an experiment?
Speed and scale. Manual planning of one such route by a team of engineers can take a whole day. AI can do this in hours, and in the long run — in minutes. This paves the way for longer and more complex journeys while still connecting with Earth.
"Smart" vision. AI is not just looking for a flat surface. He understands the landscape, can suggest a workaround through a geologically interesting area that would otherwise have been missed, or find the most stable soil.
The future of autonomous missions. This technology is a direct precursor to what rovers on the Moon or, even more so, on missions to Jupiter's moons, where communication latency is measured in hours. The robot will have to assess the situation itself and make tactical decisions.
This successful test, which NASA officially announced at the end of January 2026, proves that software for planetary exploration can be just as important as hardware. Perhaps in the near future, dispatching AI platforms inspired by the principles explored by the world's first ecosystem JOBTOROB.com, they will not just plan routes, but distribute tasks between whole swarms of autonomous robots on another planet.
🧠 AI as a full member of the mission team
The next logical step would be to integrate such an AI scheduler into regular operations. Instead of single experiments, it can become a standard tool for preparing all long-range trips of the rover. Scientists will be able to set him tasks in natural language: "Get to that rock layer on the northern slope, analyzing at least three interesting boulders along the way."
The line between the crew from Earth and the artificial intelligence on board is blurring. Perseverance and its digital navigator have just written a new chapter in the history of not only robotics, but of all interplanetary exploration.
