AI That Asks Questions Teaches Kids Better Climate Decision-Making

Picture a classroom deciding how a town should cut carbon emissions. They must weigh science, cost, and fairness. A researcher tested whether an AI conversation partner helps teenagers make such decisions. In the study, 270 secondary students were split into three groups: one used an AI that asked them questions during research, one did the same research but without AI, and one got traditional instruction. All started at similar, fairly low skill levels.

After the lessons, all three groups improved, but not equally. The AI-assisted group made the biggest gains, clearly ahead of the no-AI inquiry group and far ahead of traditional teaching. The biggest advantages showed up in skills like considering stakeholders, weighing alternatives, planning implementation, and monitoring outcomes. These are exactly the messy, real-world parts of decision-making that are hardest to teach.

The AI seemed to help most when it acted as a coach rather than an answer machine—it questioned students, asked for evidence, and prompted them to consider other viewpoints. The number of times students checked the AI's claims against the provided sources predicted how much they improved. In other words, students who treated the AI as a partner to verify, rather than an oracle to trust, learned the most. Notably, no student in the study was flagged for over-reliance on the AI.

This matters because AI is entering classrooms fast. The result is a useful design principle: build AI tools that challenge and question, just like a good teacher, and embed them in structured inquiry. If done well, AI could help students tackle tough societal issues like climate change without outsourcing their thinking. But the study also implies that context matters—an AI alone isn't magic; it performs best when paired with good assignments and habits of checking facts.