Teacher librarian Carlos Pinheiro prepared a session for upper secondary students at the Leal da Câmara School Cluster (Portugal), focused on what it means to do schoolwork when AI is available, and how to avoid losing authorship of your own work!
In a 90-minute session held in the library, the class went through every stage of a school project, from the research question to the presentation, and learned to tell apart the uses of generative AI that support learning from those that replace it. Along the way, students audited AI answers, discussed authorship cases, wrote an AI-use statement and created their own “5 golden rules” for doing schoolwork in a world with AI.
The session was designed around 4 main moments, guiding students from a topic to the presentation of a piece of work in a world with AI, over 90 minutes.
The challenge (0–5 min)
Teacher Carlos Pinheiro began the lesson with a provocative question: “If AI can do the work, what does it mean to do it?”
The teacher presented the following scenario: Imagine you have to hand in a 10-page assignment tomorrow.
A: you spend hours researching, selecting, writing and revising.
B: you ask AI to do the work and hand in the result.
Which of the two assignments taught you something? Students voted and briefly justified their choice.
The transition to the next moment was made with the question: If the goal of school is to learn, is handing in a good document enough?
A piece of work is not a document (5–12 min)
The teacher visually presented the stages of a piece of work:
THINK → QUESTION → RESEARCH → SELECT → ANALYSE → PRODUCE → VERIFY → PRESENT
He explained the difference between the learning process and the final product and, in a mini activity, presented three situations, asking which showed the most learning:
- an excellent text, but the student cannot explain it;
- a less perfect text, but the student can justify their choices;
- a text produced almost entirely by AI and presented as the student’s own work.
Key concept: the document is the result of a process; it is not the process itself.
From topic to question (12–22 min)
The teacher presented four topics: Artificial Intelligence; social media; climate change; World War II. Each pair chose a topic and turned it into a research question.
Example:
Topic: Artificial Intelligence
Sub-topic: Artificial Intelligence in education
What are the advantages of AI in education?
How can the use of generative AI change the learning process of upper secondary students?
Conclusion: a topic tells us what to investigate. A question tells us what we want to find out.
Researching is not asking AI (22–32 min)
The teacher presented a short AI answer containing, for example, a statistic, a date, a quotation, a researcher’s name and a bibliographic reference.
He first asked: What in this answer can we accept immediately? And then: What do we need to verify? How? He then introduced the principle Find → Confirm → Compare.
Key message: an AI answer can be a starting point, but it does not replace consulting reliable sources.
AI Traffic Light (32–44 min)
The working groups received three options for when to use or not use generative AI to do a schoolwork:
🟥 FORBIDDEN
🟨 ALLOWED WITH RESTRICTIONS
🟩 ENCOURAGED
The teacher presented situations in turn, such as:
- Asking AI to write the whole assignment.
- Asking for suggestions of research questions.
- Using AI to improve the clarity of your own text.
- Using AI in a test without permission.
- Asking AI to criticise the arguments in a piece of work.
After each situation, students chose a colour and justified it.
Key message: the question is not just “Can I use AI?”. It is “What for, and under what conditions can I use it for this task?”
Auditing an AI answer (44–57 min)
Each group received an AI answer with a mission: “Find the problems.” Students had to identify:
- a claim that needs verification;
- incomplete information;
- a possible generalisation;
- a source they should look for;
- a question that was left unanswered.
At the end, each group presented one finding.
Goal: to move from an attitude of “AI said so” to an attitude of “I’ll verify it”.
Authorship and responsibility (57–67 min)
The teacher presented four cases:
- The student wrote the assignment and used AI to correct some mistakes.
- The student asked AI for structure suggestions, researched the sources and wrote the text.
- The student asked AI to write the assignment and changed a few sentences.
- The student handed in a text produced by AI without declaring its use.
He then asked: What distinguishes these cases? The discussion was guided by four criteria: purpose + extent of use + authorship + transparency.
Conclusion: using AI does not automatically eliminate authorship. But handing in as your own something you did not produce compromises authorship and responsibility.
Declaring the use of AI (67–75 min)
The teacher presented a simple template:
- Tool used:
- Purpose:
- How it was used:
- AI’s contribution:
- Work done by the student:
- Verification/review carried out:
Students wrote a fictional statement for a situation presented by the teacher.
Key message: declaring the use of AI makes the process visible.
Building AI-proof work (75–83 min)
The teacher presented the idea: A GOOD PIECE OF WORK LEAVES TRACES OF THINKING.
Students identified what they should keep while doing a piece of work: initial question; plan; keywords; sources consulted; notes; drafts; intermediate versions; decisions made; relevant prompts; use of AI; final revision; reflection.
Finally, the 5-minute test was introduced: If tomorrow you had to explain your work for five minutes without looking at the document, could you?
Golden rules (83-90 min)
Each group created five golden rules for doing schoolwork in a world with AI. The proposals were collected on the board and, with the class, a final version was built. A possible synthesis:
5 GOLDEN RULES
1. Think before you ask. AI does not replace your question.
2. Research before you believe. An AI answer is not a source.
3. Use AI to learn, not to disappear from your own work.
4. Verify everything you hand in. The responsibility is yours.
5. If you use AI, declare it. Transparency is also a skill.
Final message of the lesson: “AI can help you do better work. But learning, deciding, verifying and taking authorship remain your tasks.” Teacher Carlos Pinheiro
Assessment and conclusion
Assessment was mainly formative and each activity generated evidence of learning: the research question (Moment 1), the AI traffic light and the AI answer audit (Moment 2), the case analysis, the declaration exercise and the 5-minute test (Moment 3) and the 5 golden rules (Moment 4).
In the final minutes, each student could also complete an exit ticket:
- Before today, I thought that…
- Now I understand that…
- In my next piece of work, I will…
This short record made it possible to see the change in understanding and reinforced the metacognitive dimension of the session.
Conclusion
In 90 minutes, students understood that the document is the result of a process, not the process itself, and moved from an attitude of “AI said so” to one of “I’ll verify it”. Using AI does not automatically eliminate authorship, but it requires transparency: declaring the use of AI makes the process visible, and a good piece of work leaves traces of the thinking of the person who made it.
Teacher’s Reflections
“Holding these sessions has reinforced a belief that I consider fundamental: students are aware of the impact that Artificial Intelligence is having on their lives and at school. They recognize the possibilities of these tools, use them, experiment with them, and are beginning to realize that they are profoundly transforming the way they learn, research, create, and communicate. This awareness, which I have observed in the various classes I have worked with, serves as an important starting point for any educational intervention.
One of the main conclusions I draw from these activities with students is that school assignments can no longer follow the same paradigm as before the advent of generative AI. It makes no sense to continue assigning tasks based on the same assumptions, as if the production of texts, the organization of information, and the creation of documents had not been profoundly altered. We need to rethink what we ask of students, what we value, and how we assess them. More than the final product, it is important to understand and value the processes, the decisions, critical thinking, the ability to question and verify information, and the effective understanding demonstrated by students.
Although national guidelines for the use of Artificial Intelligence in education are currently being developed—with a planned effective date in January—I believe that this process, while important, should not lead schools to delay taking action. Preparing students and teachers for the challenges of AI cannot be made contingent on the publication of ministerial guidelines. It is a responsibility that schools already have, and one that requires reflection, training, experimentation, and the establishment of principles for use tailored to different educational contexts.
I also want to highlight an aspect that I consider indispensable: listening to students. We cannot develop educational responses to AI without understanding their experiences, perceptions, doubts, and practices. They are the ones who are already integrating these tools into their daily lives, often before teachers and the institutions themselves have been able to keep pace with this transformation. Listening to them does not mean indiscriminately accepting all their practices, but rather recognizing that they are key partners in building a culture of critical, responsible, and mindful use of AI.
Schools should not limit themselves to establishing rules about what students can and cannot do with these tools. Above all, they should help students understand when, why, and how to use them, without losing their intellectual autonomy, creativity, and responsibility for their own work. The challenge is not to teach students to do assignments in spite of AI, but to prepare them to learn, think, and produce knowledge in a world where AI is already a part of reality.” Teacher Librarian Carlos Pinheiro
Lesson plan developed in classroom in November 2026.
