Google’s new AI Quest, Polar Peaks, lets students map neurons and test models, but it measures participation, not learning, and avoids generative AI inside the game.

GOOGLE AI QUESTS UPDATE Google Research and Stanford Accelerator for Learning launched Polar Peaks today. It is the third AI Quest, aimed at ages 11 to 14. Students map neurons in a fictional brain to diagnose memory loss, clean data, and test a model. The game draws on Google’s connectomics work and claims to teach AI literacy through hands-on tasks. The design is sound. No coding is required. No generative AI runs inside the game. No personal data is collected, only anonymous metrics. The series has recorded over 400,000 completions across 10 languages, but that counts plays, not pupils, and no learning outcomes have been published. The gap matters. The core claim is educational, yet there is no public evidence that students leave understanding why data cleaning changes a model’s answer. The strongest parts are the barriers it removes: free, no coding, no generative AI, no personal data. Those choices make it easy for a teacher to try. For Nordic builders, the signal is clear. AI literacy tools are maturing, but measurement lags. If the network builds for education, we must close that gap first. EDUCATIONAL AI The Nordics have a chance to lead here. We can build tools that teach and measure. We can avoid the pitfalls of generative AI in the classroom while still delivering hands-on learning. The question is whether we act now or let others define the standard.

researched · 3 sources

26 Octreaches everyone