The use of generative artificial intelligence as support for learning mathematics is becoming increasingly common among high school students, especially in situations where individual teacher support is not available. A study conducted with 330 students in China shows that students have clear preferences about what such support should look like.
Students want help, but they do not want someone to solve tasks for them. Most students prefer hints and guidance that help them reach a solution on their own, while only a small percentage want AI to provide a complete solution, and researchers interpret this as a desire to maintain autonomy in learning.
Students also have clear views on the timing of intervention. They most often want help at the moment they get stuck while solving a task or when they do not understand the problem, while significantly fewer students want help when they lose focus. Proactive assistance is rated positively in terms of adaptability, but at the same time it can be somewhat distracting.
When it comes to data, students are willing to share information about how they solve a task (such as steps and mistakes), but they are much less willing to allow tracking of their attention and behavior during work.
