AI literacy activities for Geography lessons

• The Earth follows certain rules, like tectonic plates moving or magma rising. AI systems look for patterns in data to predict when an earthquake or eruption might happen.
• This is helpful for saving lives, but computers can only "know" what we tell them through data.
• AI is a powerful tool for scientists, but it is not magic. It relies on sensors (like seismometers) and historical records.

• Pattern: something that happens in a regular or repeated way.
• Prediction: using what we know now to make a smart guess about what will happen next.
• Data: information, like numbers or facts, that a computer uses to learn.
• Sensor: a tool that "feels" or "hears" things, like a machine that feels the ground shaking.

• UNESCO: Pupils demonstrated an understanding of how AI identifies geographical features through pattern recognition in satellite imagery and hand-drawn models.
• OECD AILit: Pupils engaged with AI predictions by evaluating the reliability of an earthquake warning system in a structured debate.
• PISA 2029: Pupils identified how a simple pattern-matching algorithm works by simulating a tectonic sensor in an unplugged activity.
• DfE GenAI Safety: Pupils discussed the risks of over-relying on automated warnings during the Vulcania scenario.

A short selection. A full guide maps every activity to the frameworks you choose for your region.

• In Geography, we use models all the time, like the long profile of a river. AI uses training data (past records of river levels and shapes) to predict future changes, such as where a meander might erode or a floodplain might fill.
• An AI model is only as good as the data it is fed. If a model only "sees" images of the Thames, it might not recognise a fast-flowing mountain stream as a river at all. This is called bias.
• Technology is a tool built by humans, not an infallible source of truth.

• Training data: the examples (like photos of rivers) we give a computer so it can learn what a river looks like.
• Algorithm: a set of step-by-step instructions a computer follows to solve a problem, like predicting where a river might flood.
• Bias: when an AI makes mistakes because the data we gave it was missing important information.
• Computer model: a digital version of the real world used to test what might happen if things change, like heavy rain on a floodplain.

• UNESCO: Students applied their knowledge of river landforms to train a classification model, demonstrating how AI recognises patterns in geographical data.
• OECD AILit: Pupils explored the safety and ethical dimensions of automated flood warnings, identifying why human oversight is necessary.
• PISA 2029: Students identified how simple predictive algorithms operate in environmental monitoring and acknowledged their limitations.
• Anthropic AI Fluency (4D): Students practised discernment by questioning AI flood predictions based on incomplete historical data.

A short selection. A full guide maps every activity to the frameworks you choose for your region.