AI4Mars, a citizen science project, puts users to work classifying photos of the Red Planet’s sandy, dusty surface. It was designed to help NASA’s Mars rovers identify hazardous terrain.

Participants comb through images of Mars’ surface and label sandy, soily and rocky terrain and ID big rocks. The data will be used to train neural networks that will eventually help current and future Mars rovers figure out where it’s safe to drive in a system similar to the one self-driving cars use to avoid road hazards.

Doing so could help NASA’s Curiosity and Perseverance rovers, and future Mars exploration machines, avoid the fate of Spirit, a Mars rover that got stuck in soft soil after six years of tooling around the Red Planet. Though Spirit outlasted the 90-day mission scientists had initially planned, it might have lasted far longer if not for the inhospitable Martian terrain. Instead, it was marooned, and an attempt to turn it into a stationary science platform failed.

AI4Mars is overseen by researchers from the NASA Jet Propulsion Laboratory, but the vast majority of team members are ordinary folks at home.

Thus far, about 7,400 volunteers have classified over 60,000 chunks of terrain — and you can help. Visit bit.ly/AI4Mars to join in.