Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Planning a course for NASA’s Perseverance rover140 million miles away on Mars, is much more difficult than determining a driving route here on Earth, where we can enter an address into Google Maps and be on our way in seconds. The vehicle’s path is usually planned by a team at NASA’s Jet Propulsion Laboratory to take into account terrain, obstacles and potential hazards, lest the vehicle roll over or be damaged.
For the first time, NASA’s Jet Propulsion Laboratory Uses artificial intelligence To plan a training course for perseverance, and it seems to have worked.
The two demonstrations, which took place on December 8 and 10, were previously planned Anthropic Claude Artificial Intelligence The models were double-checked by JPL to ensure that the artificial intelligence did not accidentally push the vehicle into a ditch. I drove the Perseverance just under 1,500 feet across both drives with no documented problems.
NASA took a similar approach to plotting waypoints as it does with human operators. Claude was fed the same satellite images and data from NASA’s Mars Reconnaissance Orbiter that JPL scientists would use, and then asked to plot waypoints that Perseverance could safely handle.
The resulting trajectory was modified slightly by NASA and then shipped to Perseverance, which then drove the trajectory independently.
“This demonstration shows how advanced our capabilities are and expands the way we will explore other worlds,” said NASA Administrator Jared Isaacman. “Autonomous technologies like this can help missions operate more efficiently, respond to challenging terrain and increase science yield as distance from Earth increases. It’s a powerful example of teams carefully and responsibly applying new technology to real-world operations.”
You can watch the flight on December 10 on NASA’s YouTube channel, where it has been condensed 52 second video.
The path planned by Claude is shown in purple, and the actual path followed in orange. NASA scientists only had to make minor adjustments to the AI’s trajectory.
A more efficient way to do this
While artificial intelligence is largely known as Ramp providerwhich was quickly blamed Impairing people’s online experienceIt can be useful in some scientific endeavors. It takes years of time to analyze images and data, map out persistence waypoints, and then implement them.
According to NASA, waypoints are typically not set closer than 330 feet away, which means Perseverance is exploring the Red Planet one football field at a time. It sets out on its epic climb out of Jezero Crater in 2024. The journey took Perseverance 3.5 months, and in all, the rover climbed a total of 1,640 vertical feet. As of December 2025, the vehicle has traveled a total of only 25 miles in about four years.
The goal, according to space roboticist Vande Verma, is to allow Perseverance (and… Other Mars rovers) Travel further with “reduced operator workload.”
Verma also points out that AI could be used to identify interesting features on the planet, saving humanities teams time by eliminating the need to manually verify “huge amounts of rover imagery.”
“This demonstration shows how advanced our capabilities are and expands the way we will explore other worlds,” Isaacman said. “Autonomous technologies like this can help missions operate more efficiently, respond to challenging terrain and increase science yield as distance from Earth increases. It’s a powerful example of teams carefully and responsibly applying new technology to real-world operations.”