In June 2026, the OECD and the European Commission released a new AI literacy framework for primary and secondary schools, and it is being applauded across education. This episode looks at what the applause skips.

The framework feeds the international assessment that shapes national curricula through the end of the decade, and it names what a generation should be able to do with AI. What it does not do is teach the method. It tells students to check their own work and set their own checkpoints, and it never supplies a procedure or a standard for how. Sixty years of learning science places durable learning in the method of use, in deliberate friction and in support that is withdrawn so the learner takes over the capability. Strip that out and checking the work becomes trusting the tool.

The same gap runs through accountability. In the framework, a checkpoint is a student monitoring herself, with no named adult and no one who answers when the AI is wrong. Regulators already define a real checkpoint as a specific person with the authority to override the system and the responsibility for what follows. A framework preparing a generation for AI-mediated work teaches a weaker standard than a bank already owes a loan applicant.

The episode closes on the cost, which will not show in a 2031 test score. It will show in a generation taught that a checkpoint is something they set for themselves, arriving in workplaces that ask who answers when the AI output ships and is wrong.

For the full analysis and all sources, visit https://basilpuglisi.com/ailit-framework-rule-optional/

This is an AI-generated audio overview produced with NotebookLM. Content may include inaccuracies; the written article is the source of record. #AIassisted using HAIA Ecosystem