For teachers

This course does not just warn about artificial intelligence. It trains judgment, through a repeatable method you can run in any subject.

Why it is different

• It names the real risk: practice debt. Letting the machine take your practice is a loan that accrues interest; the ability does not settle in the student. The course is built to prevent it.
• Every task starts with the student's own work (the human anchor), then lets the machine amplify it, and ends in accountability (the accountability document). A student cannot finish by prompting alone.
• The same three-step spine across all 21 modules. Learn one, you know them all.
• Funded by the Development Fund for Educational Materials (Rannís) and free to use (CC BY-SA 4.0).

The four competency lines

Each module is tagged with one of four competency lines. This is the project's own internal classification, the teacher's map behind the student-facing surface. The learning outcomes themselves, in the sense of the national curriculum, are in a separate document linked below.

• ◆Substrate judgment

  Telling where machine output is trustworthy, where it is fragile, and when primary sources or human experience are needed.

• →Direction-holding

  Holding your own goal instead of letting the tool steer the task toward the next comfortable answer.

• ◎Collaborative action

  Using the machine as a colleague, not a substitute. Delegate, review, reject, correct, and stay accountable.

• △Edge literacy

  Practicing work where neither teacher nor student has a perfect answer in advance.

National curriculum: learning outcomes, key competencies, levels

The course is mapped onto the Icelandic national curriculum for upper secondary schools: ten learning outcomes, a mapping onto the nine key competency areas, level placement, and the case for the three-step method as formative assessment.

• Ten learning outcomes in national-curriculum form, shared across all modules, because they all follow the same spine.
• Three key competency areas at the core: numeracy and information literacy, learning capability, and creative thinking.
• Designed for level 2. The foundations sit at level 1 with no prerequisites. Modules 05, 09 and 20 can run at level 3.
• The three-step method is formative assessment: the process is graded, not the artifact, and the highest mark is earned for a reasoned rejection rather than a polished document.
• The full mapping of learning outcomes and key competencies, with the credit calculation, comes with the teacher pack as a separate document.

How to run it

Start with the foundations, then pick modules by your class's interest. Each module is self-contained, takes one or two lessons, and follows the same three steps. Students may use any AI tool; a free account is enough, and no task requires personal data.

• Step 1, the human anchor, should happen WITHOUT a machine. The material says so: the step does not begin at the keyboard. In class the simplest approach is to hand it out on paper or set laptops to airplane mode.
• The exercise needs no AI access. The response the student critiques is prepared in advance, so one device serves two students. The claims are present in the page even with JavaScript disabled, so a blank exercise prints as a worksheet.
• Only step 2, the algorithmic pivot, actually requires a machine. Step 3 needs a device to record answers and nothing more.
• A realistic split for a 40 minute lesson: 5 to introduce, 20 to 25 on the anchor on paper, 10 to read out and compare. A full three step run takes two lessons.
• The 1,000 word requirement for the anchor is homework rather than classwork. The material allows a shorter test version and 150 to 250 words is enough in class.