NASA’s Perseverance Mars rover has marked a historic milestone by completing its first drives on the Red Planet that were planned entirely by artificial intelligence instead of human route planners. In early December 2025, the rover executed two autonomous drives across Jezero Crater’s rugged terrain, following waypoints generated by a generative AI system developed in collaboration with NASA’s Jet Propulsion Laboratory (JPL) and Anthropic. This represents a significant evolution in robotic space exploration, as AI now handles complex decision-making tasks on another world.
Traditionally, rover navigation planners on Earth analyze terrain imagery and data, then send detailed waypoints to Mars with many hours of delay due to communication lag. However, for these December drives, vision-capable AI models processed high-resolution orbital and surface data to identify hazards, slopes, sand ripples, and rock formations, and to generate safe routes for Perseverance to follow — a task usually managed manually by mission engineers.
The AI-planned drives occurred on the 1,707th and 1,709th Martian days of the mission, with the rover traveling hundreds of meters on each occasion using its own autonomous systems to execute movement based on the AI’s guidance. This early demonstration suggests that future Mars rovers could travel longer distances more efficiently and respond to unexpected surface challenges without waiting for detailed human planning back on Earth.
NASA scientists see this achievement as a major step toward more efficient extraterrestrial exploration, especially as missions reach farther destinations where communication delays make real-time control impossible. By integrating AI into navigation planning, teams hope to reduce workload, increase scientific return, and prepare for future missions that may rely on autonomy to explore distant and challenging environments on Mars and beyond.
