National Guidelines 2026: what really changes for teachers (upper secondary and university)
TheNational Guidelines 2026push toward an idea of school (and tertiary education) centered on transferable skills, responsibility in the use of sources, and continuity across fields of knowledge. For upper-secondary teachers and for those working inAI in university teaching, the message is clear: innovation is not an “add-on module,” but a cross-cutting criterion for design, assessment, and inclusion.
Three priorities emerge repeatedly: enhancing classical knowledge (reading, writing, argumentation), strengthening scientific-technological skills (with attention toAI and STEM disciplines) and developing a digital culture that includes the conscious use ofartificial intelligence tools at school. This has concrete implications: recalibrating learning objectives (not only “knowing,” but “being able to do and being able to justify”), making product quality criteria explicit, and making the role of AI in the study and production process transparent.
For teachers, instructional responsibility expands: it is not enough to “ban” or “allow” AI. You need to design tasks that require decision-making, traceability, comparison across sources, and metacognitive reflection. In this sense, AI becomes an authentic context for training skills: checking reliability, recognizing bias, documenting choices, citing correctly, and protecting personal data.
Integrating AI into the curriculum: a 4-step operational model (without overturning the programs)
StudierAIcan help teachers and students turn content into learning activities and routines: verifiable summaries, flashcards, quizzes, oral simulations, and planners. The idea is not to delegate, but tomake the process visible
1)In practice, you can use these elements in a way that is consistent with integrating AI into the curriculum: (1) generate comprehension and application questions from course materials; (2) have students prepare an oral simulation with constraints (time, disciplinary vocabulary, examples); (3) monitor critical points with short, repeated quizzes; (4) ask for a brief “method note” on how the student used AI and what checks they performed. If you want to try it with your class or for a university course, you canstart for free
who we are.: ask for a product that has an audience and a purpose (report, podcast, scientific poster, policy brief, argumentative commentary). Include a mandatory step comparing the AI output with disciplinary sources (textbook, articles, data).
3)Select tools and resources: choose a fewAI tools for teachersand students, defining what to use them for (e.g., generating questions, simulating an oral interview, proposing alternatives). Provide equivalent “non-AI” resources for inclusion and accessibility (guided outlines, glossaries, prompts).
4)Assess process and quality: build clear criteria (rubric) and ask for evidence: prompts used, successive versions, source-check notes, rationales for choices. Here AI becomes a “lab” for observing method, not a shortcut to the product.
To make the model sustainable, prepare a standard “task sheet”: objective, instructions, what is allowed with AI, what must be declared, assessment criteria, and a minimal citation example (including AI, when relevant).
Examples by subject: how to use AI meaningfully (humanities, STEM, languages, social sciences)
The goal is not to “make students use AI,” but to build activities in which AI makes reasoning, errors, and revisions visible. Below are some replicable ideas, with attention tolimits, biasand the conscious use of sources.
- Humanities (Italian/history/philosophy): critical comparison between two AI-generated summaries and a textbook chapter. Students highlight omissions, anachronisms, “slanted” lexical choices, and rewrite an argued version with notes and citations.
- STEM (mathematics/physics/computer science): analysis of AI-produced solutions to a problem. Task: identify unjustified steps, verify with their own calculations or simulations, propose a correction, and explain why the error is plausible (e.g., unit confusion, implicit assumptions).
- Languages: role-play with AI to simulate an interview (defined level and register), then a metalinguistic debriefing: recurring errors, lexical alternatives, pragmatics and politeness. Assessment focused on revision and awareness, not on the first draft.
- Social sciences/law/economics: writing a policy brief on a local issue. AI proposes options and counterarguments; students must anchor each point to data and verifiable sources, indicate impacts on stakeholders, and declare value choices.
In all subjects, an effective practice is “triangulation”: AI output + primary/secondary source + human check (calculation, close reading, comparison with data). This way AI becomes an opportunity to teach method and responsibility, in line with the expectations of theNational Guidelines 2026.
Tests and assessment in the AI era: rubrics, orality, processes, and academic integrity
With AI, traditional “product-only” assessments become fragile. The answer is not to raise barriers, but to increase reliability by assessingprocess, orality, and traceability. A well-built rubric can include: source quality, accuracy, argumentative coherence, methodological choices, revision, and declaration of AI use.
Practical strategies (school and university):
- Two-stage assessments: draft (even with AI, declared) + in-class/lab revision with constraints and mandatory sources.
- Short oral interview about the submitted work: the student explains choices, critical steps, discarded alternatives, and how they verified information.
- Evidence portfolio: essential prompts, sources consulted, annotations, errors found in the AI output, improvements made.
- “Anti-copy” but pro-learning tasks: local data, texts discussed in class, questions that require application to a specific case and justification of choices.
On integrity, a simple policy is useful: what is allowed, what must be declared, how to cite, and what the consequences are for improper use. At university this connects directly toacademic integritypractices; in upper secondary it becomes education in responsibility and digital citizenship.
StudierAI as a teaching ally: design, guided study, and skills monitoring
To make AI operational without increasing workload, a platform that supports both design and study can be useful.StudierAIcan help teachers and students turn content into learning activities and routines: verifiable summaries, flashcards, quizzes, oral simulations, and planners. The idea is not to delegate, but tomake the process visible(what I studied, what I still don’t know, where I make mistakes) and link it to observable skills required by the MIM.
In practice, you can use these elements in a way that is consistent with integrating AI into the curriculum: (1) generate comprehension and application questions from course materials; (2) have students prepare an oral simulation with constraints (time, disciplinary vocabulary, examples); (3) monitor critical points with short, repeated quizzes; (4) ask for a brief “method note” on how the student used AI and what checks they performed. If you want to try it with your class or for a university course, you canstart for freeand share a common study routine. To explore the approach and the educational vision of the project in more depth, you can find more information inwho we are.
Bringing AI into everyday practices ultimately means strengthening what school and university have always demanded: rigor, clarity, responsibility, and the ability to argue. The National Guidelines 2026 provide the framework; it is up to instructional design to turn it into reliable, inclusive, and assessable learning experiences.
