Google's Polar Peaks turns brain mapping into an AI literacy lesson

The third AI Quest gives students data-cleaning and model-testing tasks while new modules help teachers bring the programme into class.

Google Research and the Stanford Accelerator for Learning have added a neuroscience-themed lesson to their AI Quests programme. Announced on October 26, Polar Peaks asks students aged 11 to 14 to take on the role of researchers mapping neurons while investigating a game character's memory loss.

The activities include cleaning data and testing models, with an emphasis on critical thinking and human oversight. Google describes the quest as inspired by its connectomics research. It follows earlier lessons drawing on flood forecasting and eye-disease detection, giving students another example of AI applied to a scientific problem.

Google says more than 400,000 quests have been completed across the programme, which is localised into ten languages. That is a completion count across the series, rather than a count of individual students or an evaluation of learning gains. Teacher training for the first two quests is available, with the third lesson's module planned for the following month.

The educational opportunity lies in showing the steps around a model's answer. Cleaning an input or testing a prediction gives students a way to see why an AI result depends on data and evaluation. It also makes the role of a human investigator concrete.

The storyline gives that process a concrete purpose: students scan neurons with in-game mentors to investigate memory loss. After completing the quest, they hear from researchers involved in real connectomics work, including mapping a fruit-fly brain. The progression connects a simplified classroom challenge with the research that inspired it, while keeping the game distinct from a clinical diagnostic tool.

Sources

- [Our newest AI Quest: Mapping the human brain](https://blog.google/products-and-platforms/products/education/ai-quests-polar-peaks/) — Google