Generative AI in Education – Building a Common Culture – Edition 2026

The document offers a very operational framework for educational teams (primary and secondary) wishing to frame and share the uses of generative AI: training, charter, governance, co-education, ready-to-use activities. It is particularly useful for headteachers, inspectors, digital referents, school life coordinators and teachers who need to move from individual experiments to a coherent school policy. The numerous "activities" are directly transposable into team animations, class sessions or family time, without having to reconstruct the sequences.

Source: 📒 Generative AI in Education – Building a Common Culture – Edition 2026 – Guidelines, resources and activities for the classroom

 📜🔗LINK

Number of pages: 31

1. ANALYTICAL SUMMARY

Context, audiences and issues

The document addresses the entire educational community (teachers, educational staff, headteachers, IEN, ERUN, RUPN, CPE, AESH, administrative staff) in a context of rapid generalisation of generative AIs since 2022 and the formalisation of a national usage framework published in June 2025 by the Ministry of National Education. It presents generative AIs as a "total social fact" permeating work, communication, leisure, and school life, and emphasises the need for a common foundation of understanding for all staff, regardless of individual use in the classroom. The document articulates four major challenges: systemic training for staff, co-education with families, integration of AI into the school project, and civic education for students (work, inequalities, misinformation, environment, sovereignty).

Operational contributions for the field

The document proposes a gradual pathway: training schemes (European, national, academic, local), needs assessment activities, tools to co-construct the AI usage charter, co-education scenarios with families, mapping of resource actors, communities of practice, and "ready-to-use" educational activities from primary to secondary school. It provides detailed examples of workshops (mapping training needs, AI family evenings, AI club, digital sobriety audit, debates on work and deepfakes), all scripted (objectives, duration, phases) and directly reusable. Finally, it explicitly links these practices to national schemes (Pix AI, National Observatory of Practices with AI, Compar:IA, CREIA, CNIL, national AI strategy) to anchor local actions within a broader regulatory and political framework.

2. KEY POINTS OF THE DOCUMENT

1. Common culture of AI for all staff (p. 4–7)

   The document argues that the central question is not "should I use AI in the classroom?", but "am I able to support students in understanding these tools?". It emphasises the training of all staff (teachers, administrative staff, AESH, CPE, psychologists, managers) as a condition for educational coherence, relying on the national framework for the use of AI in education published in June 2025.

2. Structured training ecosystem, from the MOOC AI4T to Pix IA (p. 7–9)

   The document maps a multi-level ecosystem: European MOOC AI4T, CREIA, Campus of AI and Digital Innovation, CanoTech, Pix+ Edu, Magistère, EAFC, DRANE, Canopé Workshops, district and academic training plans. It details the deployment of two mandatory Pix IA courses for students in Year 10, Year 11, and Year 12 CAP starting from BO No. 6 of 5 February 2026, as a common foundation for the teams.

3. Co-education with families and dialogue tools (p. 10–12)

   The document dedicates a section to family involvement: accessible information on what AI can/cannot do, various formats (meetings, ENT, summary documents, web pages, AI Cafés), and highlighting external stakeholders (libraries, community centres, associations, local authorities). It proposes structured activities to prepare a family evening on AI and to organise a collection of students' voices regarding their uses, fears, and expectations.

4. Establishment charter and shared governance of AI (p. 13–17)

   The text makes the AI charter a vector for appropriating the national framework rather than just a compliance document, emphasising co-construction with teachers, non-teaching staff, AESH, families, and students. It describes the role of leadership (management, IEN, RUPN, ERUN, DRANE, Ateliers Canopé) in integrating AI into the school project and coordinating training, charter, co-education, and support for professional use (GDPR, DPO, CNIL).

5. Citizen dimension: work, inequalities, environment, sovereignty (pp. 20–28)

   The file offers a structured analysis of the social, environmental, and geopolitical issues of generative AIs: transformations in work, potential widening of access and skills inequalities, hyper-manipulations, environmental footprint, economic concentration, and digital sovereignty. It provides educational activities (from cycle 3 to high school) to develop critical thinking (true/false/impossible, impacts on professions, personal digital footprint, mapping of AI stakeholders) and introduces tools like Compar:IA to question the choice of models and digital sobriety from the classroom.

3. ACTION PATHS FOR LOCAL ACTORS

1. Organise a shared diagnosis of AI training needs (pp. 4–6, 7–9)

   School leaders, heads of establishments, or digital referents can use the activity "Mapping training needs as a team" to have each member take an initial quiz, then anonymously share areas of uncertainty (functioning, ethics, GDPR, sovereignty). This diagnosis is then used to build a training plan combining MOOC AI4T, CanoTech modules, Magistère, EAFC/DSDEN training, and support from Ateliers Canopé, distinguishing between individual and collective needs.

2. Incorporate generative AI into the establishment project via a living charter (pp. 13–15, 28–29)

   Teams can start from the national framework, the internal regulations, and the existing IT charter to develop a unique digital/AI charter, rather than adding an additional document. The proposed method includes preparation by a small group, a pedagogical advisory session to discuss actual uses, points of tension (personal data, unreferenced tools), joint drafting of guidelines, and an annual review cycle.

3. Deploy co-education actions on AI with families and local stakeholders (pp. 10–12)

   Establishments can organise AI Cafés or family evenings by relying on the Café AI resources and the local Canopé Workshop (support for design, facilitation, materials). The activity 'Preparing a family evening on AI' provides a precise framework: choice of format, selection of concrete situations from classroom life, preparation of a short support document and a summary document given to parents, subsequently shared via the ENT or the establishment's website.

4. Create and animate AI practice communities (adults and students) (pp. 18–19)

   To avoid isolating innovators, the file recommends institutionalising peer exchange times (team meetings, mutualisation workshops, capitalisation sheets) and linking them to the National Observatory of pedagogical practices with AI and the TraAM. For students, the creation of AI clubs in middle/high schools, reflection groups around the charter, or EMI workshops allows for the development of agency, critical thinking, and participation in decisions regarding digital education.

5. Integrate digital sobriety and sovereignty into AI actions (pp. 23–27)

   Teams can use the activities of "audit of the institution's digital usage" and "My digital footprint" to co-construct recommendations for sobriety (choosing more frugal models, limiting certain usages, improving the quality of prompts to reduce requests). At the same time, mapping global AI stakeholders and working on digital sovereignty allows for clarifying with students and staff the issues of technological dependence and the choice of tools compliant with GDPR and national guidelines.

4. ADDITIONAL REFERENCES 

🔍➕ For more information, see the articles referenced by "Practices in Health" on the theme of digital and AI ➡️🔗https://pratiquesensante.odoo.com/2-6-intelligence-artificielle-numerique

5. CROSS-ANALYSIS — VALUES OF PRACTICES IN HEALTH

• Literacy: the document provides clear explanations, unplugged activities, and simplified materials (family evenings, one-page documents) for different levels of understanding, including non-experts.

• Empowerment: students are explicitly invited to analyse their usage, co-write the charter, run AI clubs, and present proposals in CVC/CVL, which strengthens their ability to act on their digital environment.

• Participation: the co-construction of the charter, surveys of students, and AI Cafés open to families structure formal mechanisms for co-constructing rules and practices.

• Community health: the collective dimension is omnipresent (common culture, communities of practice, articulation between institutions, families, and socio-cultural actors) with explicit attention to inequalities in access and digital skills.

• Ethics: the text identifies biases, risks of misinformation, issues of personal data, sovereignty, and environmental impact, and refers to the CNIL and the national framework to address these issues.

• Human rights: the approach emphasises equitable access to understanding AI (including for students without equipment) and the protection of students' data through GDPR and sovereign AI.

• Intersectorality: the document recommends working with media libraries, social centres, popular education associations, local authorities, and national initiatives (CREIA, Observatory, France 2030).

• Partnership: it formalises models of collaboration between Ateliers Canopé, DRANE, EAFC, DSDEN, educational institutions, families, and cultural actors, notably through AI Cafés and training in institutions.

• Combating discrimination: the file addresses model biases, the risks of reproducing stereotypes, inequalities of access and cultural capital, and promotes a non-judgmental stance and gradual support for families and students.

6. EVALUATION OF THE RELIABILITY OF THE RESOURCE

• Scientific relevance: the resource is published by Réseau Canopé (a national public institution), relies on the national AI framework, the CRCN, the GDPR, the work of the CNIL, the National Digital Council, France 2030, AI4T, and refers to synthetic analyses on environmental impact (Roussilhe). The perspective is primarily educational and public policy-oriented, but grounded in updated institutional references (2025–2026) and cautious regarding environmental figures (emphasis on trends rather than absolute values).

• Operational relevance: the document is strongly action-oriented, with numerous detailed activities (objectives, duration, target audiences, materials, session phases) covering team training, co-education, digital education, digital sobriety, and governance, reusable without heavy rewriting. It provides links to concrete devices (Pix IA, Café IA, Observatory, Compar:IA, Pix+ Édu, Magistère), making it an immediately usable resource for teams.

7. MCQ — 5 QUESTIONS

PART 1 — MCQ (without answers)

Question 1 (pp. 13–15):

In the construction of an AI usage charter at the level of an institution, which principle is highlighted in the document?

a) The charter must be written solely by the management and then imposed on the teams.

b) The charter must be co-constructed with the entire educational community, including students and families.

c) The charter must be limited to student uses and ignore professional uses.

d) The charter must be identical in all institutions within an academy.

Question 2 (pp. 7–9, 31):

What role is assigned to the Pix IA pathways deployed from 2026 for students in Year 9, Year 10, and Year 11 CAP?

a) To completely replace the work of teachers on generative AI.

b) To constitute a common base of knowledge on AI, to be extended in class.

c) To teach students to circumvent the rules of AI usage at school.

d) To serve solely to certify high technical skills in AI.

Question 3 (pp. 23–24, 31):

What explicit recommendation does the national framework make regarding the environmental footprint of generative AIs, as outlined in the document?

a) Encourage the maximum use of AIs to optimise all educational processes

b) Consider the environmental footprint as marginal and non-priority

c) Prefer a frugal use of AI, mobilising it only if no less environmentally costly solution is suitable

d) Prohibit the use of any generative AI in educational institutions

Question 4 (p. 20–22, 31):

What does the document say about the relationship between generative AI and transformations in work?

a) Studies converge to say that AI will have no impact on jobs

b) Generative AIs shift the boundary between automatable and non-automatable tasks, including for cognitive activities

c) Judgment and critical evaluation skills are fully automatable in the short term

d) The school has no role to play in the face of these transformations

Question 5 (p. 25–27, 31):

How does the document link digital sovereignty and civic education?

a) By stating that digital sovereignty is solely a governmental matter

b) By avoiding discussion of private companies that develop AIs

c) By inviting the mapping of AI actors, analysing their legislation, and encouraging students and staff to reflect on their tool choices

d) By recommending the exclusive use of sovereign AIs already available in all institutions

PART 2 — COMMENTED CORRECTION

Question 1:

✅ Correct answer: b) The charter must be co-constructed with the entire educational community, including students and families.

📝 Explanation: the document emphasises that a charter only has value if it is embraced by those who live it, and recommends a collective development involving teachers, non-teaching staff, AESH, parents, and students, based on the national framework and actual practices. Source: pp. 13–15.

Question 2:

✅ Correct answer: b) Establish a common foundation of knowledge on AI, to be extended in the classroom.

📝 Explanation: the two Pix AI pathways cover the functioning of generative AIs, writing instructions, data protection, and environmental impacts, and are presented as a common basis for coordinating pedagogical approaches, not as a substitute for teachers' work. Source: pp. 7–9, 31.

Question 3:

✅ Correct answer: c) Prioritise a frugal use of AI, using it only if no less environmentally costly solution is suitable.

📝 Explanation: the document explicitly reiterates the principle of the national framework: AI should only be used if less resource-consuming solutions do not satisfactorily meet the need, integrating it into education for sustainable development and digital sobriety. Source: pp. 23–24, 31.

Question 4:

✅ Correct answer: b) Generative AIs shift the boundary between automatable and non-automatable tasks, including for cognitive activities.

📝 Explanation: the text highlights that tasks once considered non-routine (writing, synthesis, argumentation) are becoming partially automatable, which revalues the skills of judgement, contextualisation, and critical evaluation that schools must strengthen. Source: pp. 20–22, 31.

Question 5:

✅ Correct answer: c) By inviting the mapping of AI stakeholders, analysing their legislation, and encouraging students and staff to reflect on their choice of tools.

📝 Explanation: a section is dedicated to geopolitical issues, proposing activities to map private and public actors, analyse legislation (GDPR, AI Act), understand dependence on foreign infrastructures, and discuss the concepts of sovereign AI with students. Source: pp. 25–27, 31.

(The order of the correct answers respects the constraint: b / b / c / b / c.)

8. FAQ 

1. Who is concerned by this document within a school? (pp. 4–7, 16–17)

   The document targets the entire educational community: teachers of all subjects, educational staff (CPE, AED, psychologists), AESH, administrative staff, management (school directors, heads of establishment, IEN, IA-IPR) and digital referents (ERUN, RUPN). It emphasises the idea that understanding generative AIs is a shared responsibility, regardless of the individual use of these tools in the classroom.

2. How can I practically start an AI initiative in my school or establishment? (pp. 4–6, 7–9, 16–17)

   The document proposes starting with a collective diagnosis through a quiz and a training needs mapping activity, then building an action plan that articulates training (CanoTech, Magistère, EAFC, Pix+ Édu), a usage charter, and co-education with families. It recommends relying on the Ateliers Canopé and academic services (DRANE, DSDEN, EAFC) to obtain support and resources.

3. How to involve families in decisions about the use of AI in schools? (pp. 10–12)

   The report advocates for accessible formats: back-to-school meetings, AI Cafés, summary documents provided to parents, dedicated pages on the ENT, anchoring discussions in concrete situations (writing with AI, generated images, false but plausible answers). An activity guides the preparation of an AI family evening, including format choice, scripting examples, and possible recourse to a Canopé Workshop for facilitation.

4. How to address the risks of misinformation and deepfakes with students? (pp. 20–22)

   The report suggests media education activities where students analyse various content (texts, images, audio, video) by asking the question "true, false or impossible to know?". It connects these practices to EMI resources (CanoTech, CLÉMI) and emphasises the importance of teaching verification reflexes and understanding hyper-manipulations.

5. What can be done to limit the environmental footprint of AI in my institution? (pp. 23–24)

   The text recommends conducting an audit of digital usage (for both staff and students), identifying the most energy-consuming practices, and discussing alternatives (more frugal models, reducing unnecessary queries, clarifying needs before resorting to AI). The activity "My Digital Footprint" allows students to record their usage, estimate orders of magnitude, and produce awareness materials for the school community.

6. How to secure the professional use of AI by staff in relation to GDPR? (pp. 13–15, 28)

   The document emphasises the need for a specific charter for professional use, indicating which tools can be used, under what conditions, and reminding not to input personal data into unreferenced services. It recommends support from the academic data protection officer and CNIL resources to register processing activities and verify compliance with GDPR/AI Act.

7. How to align AI initiatives with the school project and student bodies (CVC/CVL)? (pp. 28–30)

   The document suggests incorporating civic training on AI into the existing axes of the project (media education, educational digital, citizenship, co-education), without creating an additional axis, by designating referents for monitoring. It also suggests working with student representatives to bring proposals to the CVC/CVL (modification of the charter, awareness actions), enhancing student agency and participation.

9. REWRITING IN EASY TO READ LANGUAGE 

Title

AI in schools: understanding and deciding together.

Context and issues (EASY TO READ)

• Generative AI is now everywhere in life and at school.

• Students use these tools, even if adults do not always know it.

• All adults in the school must have a basic understanding of these tools.

• This concerns teachers, management, and all staff.

• The goal is to help students understand and use AI effectively.

• The school must also discuss the risks: errors, false images, data, environment.

Operational contributions (Easy to Read)

• The document provides ideas for training and team meetings.

• It explains how to create common rules about AI in the institution.

• It offers ready-to-use activities for classes and families.

• It helps organise times for exchange with parents and students.

• It shows how to talk about work, inequalities, ecology, and politics with AI.

• It provides links to official tools for learning and training.

Key points in Easy to Read

1. Common culture of AI

• Everyone at the school must have the same basic knowledge of AI.

• It is not enough for just one person to be an expert.

• Everyone should be able to explain simply what AI does and does not do.

1. Training and official tools

• There are online and in-person training sessions for staff.

• Students have mandatory Pix AI courses to learn these topics.

• The school can rely on these courses to continue the work in class.

1. Sharing with families

• Parents often have questions and sometimes fears about AI.

• The school can organise simple meetings to explain and listen.

• A short document can summarise the rules and useful resources.

1. Common rules: the charter

• The establishment can write a charter on the use of AI.

• This charter must be written with staff, students, and parents.

• It should be reviewed regularly as tools evolve quickly.

1. Societal issues

• AI changes work, inequalities, and information.

• AI also consumes a lot of energy and resources.

• Students must learn to verify content and make responsible choices.