Could a Mars Rover Spot Safe Ground? AICAMP’s Sydney Holiday Camp

On 30 September, 17 children joined AICAMP’s AI + Space Holiday Camp at Green Square Library in Sydney. Their mission was to investigate a mysterious signal from Mars and help a rover work out which ground was safe to cross.

The challenge began with a set of images. The students looked for clues, made their own rules and then tested those rules against new examples. They soon saw how a system trained on very similar images could develop a blind spot. If it had only seen pale sand as “safe”, for instance, it might mistake another pale surface for safe ground—even when it wasn’t.

The students then tried an Arduino light sensor. It followed a simple rule: bright light meant “safe” and dark meant “unsafe”. The sensor could make an automatic decision, but it wasn’t AI because it couldn’t learn from examples. This gave the group a chance to talk about the difference between automation and AI—and why brightness alone isn’t enough to judge whether ground is safe.

After lunch, the teams tackled a grid-based route challenge. They planned a path for the rover, recorded where it got stuck or took a longer route, and compared different ways of judging the terrain. They then trained an image classifier using Teachable Machine and tested it with pictures that hadn’t been used during training. By adding a wider range of examples, they could see how the data given to an AI system affects its decisions.

At the end of the camp, each team gave a short presentation about its rover: what it was called, what blind spot the team had found, how they tried to improve the training data, and when a person should check the AI’s decision. Students recorded their work in mission logs and data sheets, and each received a certificate for completing the mission.

One idea came up throughout the day: AI can only learn from the examples it has been shown. Understanding AI means asking not only what it can do, but also what it has seen—and what it might miss.

AICAMP creates themed, hands-on learning experiences to help children and young people explore how AI works. Our activities invite students to experiment, create and work together, while thinking about how AI can make mistakes and how we can use it responsibly.

Thank you to the 17 families who joined us at Green Square Library, and to the teaching team who guided and supported the mission. We’re looking forward to welcoming students to our next AICAMP activity and setting off on a new exploration together.