NASA’s Perseverance Mars rover has completed the first drives on another planet that were planned entirely by artificial intelligence, marking a major milestone in autonomous space exploration. In a demonstration conducted on December 8 and 10, 2025, the rover followed routes generated by a vision-capable AI rather than human mission planners on Earth — a task that involves analysing terrain, identifying potential hazards, and selecting safe waypoints for traversal across the Martian surface.
For decades, rover navigation on Mars has relied on human teams reviewing imagery and terrain data to plot driving paths, then sending instructions to the rover via NASA’s Deep Space Network. The new AI approach uses generative vision-language models to interpret the same high-resolution imagery and elevation data typically used by human planners, automatically generating a continuous path that avoids risks like boulders, sand ripples and steep slopes.
The demonstration involved two successful drives: on December 8, 2025, Perseverance drove about 210 meters (689 feet), and on December 10, it covered roughly 246 meters (807 feet) using the AI-generated plan. Before commanding the rover to move, engineers validated the AI solution in a digital twin — a detailed virtual replica of Perseverance — checking hundreds of thousands of variables to ensure compatibility and safety.
NASA scientists say this breakthrough could broaden how planetary missions operate by reducing reliance on Earth-based planning and allowing rovers to navigate more autonomously, especially where communication delays are long. As missions venture farther into the solar system, such autonomous capabilities could increase efficiency, support more ambitious exploration goals, and flag scientifically interesting features for researchers.
