AI & Education: Massive Deployment and New Challenges - October 2025

Analysis of major developments in educational AI in France and internationally: mandatory PIX AI, sovereign project calls, pedagogical tools, and ethical integration challenges.

8 October 2025 Newsletter IA & Éducation

October 2025

Mandatory PIX AI Sept. 2025
AI Project Call €20M
Moodle AI Edition 12 univ.
National Strategy €560M
42 consortiums for sovereign teacher AI
15 pilot institutions
€560M national AI strategy budget

Editorial

From mandatory PIX AI deployment to sovereign project calls and international codes of conduct, the period from August 25 to October 8, 2025 marks a decisive acceleration in the integration of artificial intelligence in French education.

This newsletter reveals how the French education ecosystem is entering a phase of systemic AI integration, with record investments, widespread mandatory training, and the emergence of sophisticated pedagogical tools. But this transformation also brings new ethical and environmental challenges that question the sustainability of this digital revolution.

Key highlights: the PIX AI pathway becomes mandatory for all 8th graders, 10th graders, and first-year vocational students, while 42 consortiums compete to develop sovereign AI dedicated to teachers. Meanwhile, pedagogical tools multiply with Moodle 4.5 "AI Edition" and AI tool experimentation in French education.

Internationally, the UN creates a scientific group dedicated to educational AI and UNESCO launches its "AI-Safe School" label, signaling worldwide awareness of the issues related to this major technological transformation.

🇫🇷 France: Frameworks & Policies

PIX AI: Widespread Mandatory Training

August 25, 2025

From this school year, all 8th graders, 10th graders, and first-year vocational students must complete an AI training pathway on the PIX platform. This personalized 30-minute to 1.5-hour program covers prompting basics, how generative AI works, data management, and environmental impacts.

PIX AI Pathway Content:

  • Technical foundations: Understanding how generative AI works
  • Effective prompting: Query formulation techniques
  • Data management: Privacy and information protection
  • Environmental impact: AI carbon footprint

Also open to middle school students, high school students, and volunteer teachers, this initiative aims to democratize AI literacy throughout the education system.

Source: Banque des Territoires

Sovereign AI for Teachers: 42 Consortium Candidates

September 30, 2025

The closure of the "Sovereign AI for Teachers" project call, funded by France 2030 at €20 million, attracted 42 consortiums. The goal: develop an open and scalable generative AI to help teachers prepare lessons, grade assignments, and answer student questions.

42
candidate consortiums
€20M
France 2030 budget
2026
expected delivery

This initiative is part of the French digital sovereignty strategy, aimed at reducing dependence on American solutions in education.

New Circular "Responsible Use of Generative AI"

September 9, 2025

Each institution must now designate an AI coordinator receiving 2-day training. The circular formally prohibits the creation of personal AI accounts for minors under 15 and requires a national inventory of validated tools.

Key Measures:

  • Mandatory AI coordinator in each institution
  • AI account prohibition for under-15s
  • National inventory of validated tools (publication before October 15)
  • 2-day training for all coordinators
Source: Official Bulletin Éduscol

National AI Strategy: 2018-2025 Assessment and New Phase

2025 Update

Within the framework of the national strategy for artificial intelligence, France has allocated €560 million (2022-2025 phase) then launched a third phase in February 2025 supported by France 2030. The "AI Clusters" program (€360M) aims to create a "French MIT" with 9 centers of excellence.

Quantified Objectives (phases 2 & 3):

  • Phase 2 (2022-2025): 3,500 students/year (2,000 undergrad + 1,500 masters + 200 PhDs)
  • Phase 3 (2025-2030): 100,000 people trained by 2030
  • Achievements: 81 AI laboratories, 1,000+ startups, 13,500 direct jobs
  • Awareness: 2 million French people by end of 2027
€360M
AI Clusters 2025-2030
100k
People trained/2030
9
Centers of excellence
Source: National AI Strategy - Ministry of Economy | Phase 3 - info.gouv.fr

🛠️ Tools & Pedagogical Applications

Moodle 4.5 "AI Edition"

Launched September 16, Moodle 4.5 "AI Edition" integrates a quiz creation assistant, automated feedback, and multilingual translation. This revolutionary version is being deployed in beta at 12 pilot French universities, including the Sorbonne, Rennes, and Bordeaux.

Flagship Features:

  • Automatic quiz creation from course content
  • Personalized feedback for each learner
  • Instant translation into 15 languages
  • Integrated AI plagiarism detection
Source: Moodle

AI Tool Experimentation in Education

Since September 1, the Ministry of National Education is exploring the integration of AI tools such as Microsoft Copilot Education in select pilot institutions. This experiment comes with a strict framework for protecting minors' data.

CNIL Guarantees:

  • Hosting in France of minors' data
  • Strict control of access and usage
  • Mandatory training for participating teachers
  • Continuous evaluation of pedagogical impact
Source: Microsoft Education France
15
pilot institutions
100%
GDPR compliance

Pronote AI Assistant v21.1

Since October 2, the new version of Pronote integrates commented report card generation, dropout alerts, and schedule optimization. 6,500 middle and high schools have already activated it.

4M
connected parents

Khanmigo France - Beta

Launched September 23, Khanmigo France offers conversational tutoring in mathematics and physics, GDPR-compliant. 50 pilot classes in the Lyon and Toulouse academies are testing this pedagogical assistant.

50
pilot classes

🔬 Research & Evaluation

CNESCO: "AI & Academic Success" - Mixed Results

September 26, 2025

CNESCO's meta-analysis, covering 180 international studies, reveals an average gain of +8% in academic success with AI, but highlights a +12% increase in inequalities if usage is not supervised.

Identified Benefits

  • +8% average academic success
  • Personalization of learning
  • Increased engagement among students
  • Time savings for teachers

Identified Risks

  • +12% inequalities if usage unsupervised
  • Excessive technological dependence
  • Territorial digital divide
  • Discriminatory algorithmic biases

10 Key CNESCO Recommendations:

  • • Mandatory teacher training
  • • Algorithm transparency
  • • Screen time quotas
  • • Continuous impact evaluation
  • • Territorial access equality
  • • Student data protection
  • • Critical thinking development
  • • Energy sobriety
  • • Ethical public-private cooperation
  • • Enhanced pedagogical research
Source: CNESCO

OFCE: "Teacher-AI Productivity" - Gains and Risks

September 30, 2025

The French Economic Observatory study quantifies a gain of 4 hours per week on administrative tasks thanks to AI, but identifies a risk: 18% of teachers report "preparing their lessons less."

4h
Weekly gain
Administrative tasks
82%
Satisfied teachers
AI tools
18%
Substitution effect
Reduced preparation
Source: OFCE

ANR EVAL-IA Project: Resistant Benchmarks

October 1, 2025

The ANR EVAL-IA project, funded with €4 million over 36 months, aims to create French AI-resistant benchmarks for assessment in writing, speaking, and sciences. A consortium of excellence associates ENS, INRIA, LIP6, and ONISEP.

€4M
Total budget
36
Months of research
4
Institutions
3
Target domains
Source: ANR

🌍 International Developments

UN: International Scientific Group on Educational AI

September 25, 2025

The UN creates an independent International Scientific Group on AI, composed of 40 global experts with an explicit mandate on pedagogical impacts, teacher training, and equitable access to AI tools.

Group Missions:

  • Pedagogical impacts: Scientific evaluation of AI effects on learning
  • Teacher training: International training standards
  • Equitable access: Reducing global digital divide
  • Ethics: Governance frameworks for educational AI
Source: UN News

EU: "Generative AI in Education" Code of Conduct

September 28, 2025

The European Union launches its code of conduct for generative AI in education. French signatories include: Moodle, OpenClassrooms, and Khan Academy, committing to data transparency, minor filtering, and carbon sobriety.

Transparency

Explainable algorithms, training data sources

Minor protection

Content filtering, enhanced parental controls

Carbon sobriety

Energy optimization, CO₂ compensation

Source: European Commission

UNESCO: "AI-Safe School" Label

October 2-4, 2025

UNESCO launches its "AI-Safe School" label with 12 volunteer countries, including France. This label certifies institutions meeting strict criteria for governance, privacy, equity, and pedagogical assessment.

UNESCO Label Criteria:

  • Governance: Local AI ethics committee
  • Privacy: Student data protection
  • Equity: Equal access to AI tools
  • Assessment: Measured pedagogical impact
12
volunteer countries
500+
candidate schools
2026
first labels
Source: UNESCO
"Between massive deployment and codes of conduct, French education is entering a phase of accelerated AI integration – regulated, funded, but still fragile on ethics and sustainability."
— AI & Education Newsletter Synthesis, October 2025

Assessment: Controlled Acceleration or Headlong Rush?

A Systemic Transformation

The period from August 25 to October 8, 2025 confirms French education's entry into a phase of systemic AI integration. Mandatory PIX AI, 42 consortiums for sovereign AI, and €560M from the national strategy demonstrate strong political will.

Tools multiply and grow more sophisticated: Moodle 4.5, Copilot Education, Pronote AI v21.1... The pedagogical ecosystem is transforming at unprecedented speed.

Persistent Challenges

However, CNESCO and OFCE studies highlight risks of increased inequalities (+12% if usage unsupervised) and substitution effects among teachers. The digital sobriety challenge also emerges forcefully.

The creation of the UN international scientific group and UNESCO's "AI-Safe School" label signal worldwide awareness of ethical and societal issues.

What to Watch in November 2025

PIX AI Feedback

First feedback from mandatory pathway usage

Sovereign Projects

Delivery of winning "sovereign AI" projects

National Inventory

Publication of validated AI tools

The challenge: transforming education without losing its humanity.

← Retour à la liste des newsletters