Designing AI communication and developing a linguistic framework for how AI should talk to children in a learning context, resulting in a chatbot built with Aschehoug on their school curriculum.

Getting a straight answer out of an AI is easy. Getting an answer that a twelve-year-old can actually learn from is a different problem entirely. The way a system communicates, the words it chooses, the questions it asks back, all of it shapes whether a child engages or just copies the output and moves on.

Aschehoug publishes the textbooks that Norwegian primary school students learn from every day. The content is trusted, pedagogically grounded, and written for the classroom. The question was what happens when that same content becomes the basis for a conversational AI, and what it takes to design a chatbot that genuinely supports understanding rather than replacing it.

Ask was developed in collaboration with Aschehoug as a concept for 7th graders. The idea is that students use it to work through subject matter before bringing it into classroom discussion, with Ask drawing exclusively from Aschehoug's own textbooks rather than the open web. The name is a play on Aschehoug, the Norwegian name Ask, and the act of asking a question.

The project covered AI communication research, pedagogical research, concept development, interviews with users and experts, testing, visual design, and prototyping. It was developed together with Hanne Lockertsen. As much as the final service, the output is a linguistic framework for how AI should communicate with children in a learning context, a set of principles that could travel further than this one product.