An AI-literacy hour for incoming students
In August 2026, every incoming student at Reykjavik University, about 700 across all seven departments and the college-preparation program, got a 90-minute session on AI and learning before term began.
The town-hall meeting in August was for the people who teach. The next step was for the people who are about to learn. Before term began, Anna Sigríður Islind and I met the incoming students of every department at Reykjavik University, the college-preparation program included, about 700 in all, for ninety minutes on AI and learning.
We told them at the start what the hour was not: not a set of rules, and not a sales pitch. They are the first cohort to go through a whole degree with this tool beside them, and the university has no complete answer to offer them, because nobody does. What we could offer was one idea to hold on to. AI multiplies what you can do, but it must not take over your judgment. A bike or a taxi will both get you there, and only one of them builds your legs.
Most of the hour was spent doing rather than listening. They watched a model invent a citation and then cave when pushed on a correct answer, and learned four quick ways to check. They spent twenty minutes with an AI tutor on their own phones, on a first-year topic they thought they already knew. The tutor was set up to quiz them first and explain only afterwards, so the twenty minutes tested what they actually knew rather than handing them a summary. And they placed themselves on a ladder of five stances toward AI, from consumer to professional, where each step up moves more of the responsibility onto the person holding the phone.
We left the advanced material out on purpose. Directing tools and building things can wait; owning your judgment comes first.
Whether an hour changes anything, we cannot tell yet. What it does is make sure that no student at the university can say, a year from now, that nobody told them the difference between finishing the work and digesting it.