Students use generative AI every day, but many do not notice its biases and hallucinations, or how much of their own thinking they hand over to it. Reading about AI literacy rarely changes those habits on its own, because students need to experience the risks for themselves.

Isaac Ka Chun Wan built a one hour Interactive Document activity around a set of AI literacy readings. Students answer questions as they read, test AI chatbots and image generators themselves, and practise writing better prompts instead of relying on a single quick question.

The reading connects to HalluciHunt, an app Isaac built himself, which gives students playgrounds to test AI outputs and spot hallucinations. Students also vote on how they typically use AI, compare their answers with their classmates and with the research, and discuss what cognitive offloading means for their own learning.

The activity ends with a reflection in which each student writes a personal guideline for ethical AI use that strengthens their own judgement. It works best as a task before the lecture, so the lecture can build on what students discovered, and several questions can become group tasks for more interaction.

Instead of reading passively about AI, students test its limits, see how their peers use it and leave with their own guideline for using AI with judgement.