AI relieves repetitive tasks and creates new roles (supervision, ethics, prompt engineering). Historically, every technological revolution has destroyed some jobs but created others (printing press, computing, the internet). AI can allow teachers to focus on pedagogy, relationships, and student support. In education, replacement logic (“augmented” classes with fewer teachers) is already being tested in some countries → a real threat to the profession.

AI can adapt the pace, exercises, and teaching methods to each student. However, it could also reproduce or amplify biases present in the data, leading to unequal treatment among students...

The press recently reminded us that in 2025, despite its growing deployment, AI in education is advancing without adequate safeguards: students’ work could be used to train models, and free, insecure tools are being used improperly. Teachers and creators are being dispossessed of their intellectual production.

Generative AIs exploit unpaid human labor to generate massive private profits (value capture without redistribution).

The EU AI Act (adopted in 2024) already regulates AI uses, particularly in educational and sensitive contexts.

AI could generate between +5% and +30% growth according to the European Commission and the OECD, meaning a macroeconomic benefit for society. But a handful of companies (OpenAI, Google DeepMind, Anthropic, etc.) control the technology → concentration of power, an oligarchic dynamic. The centralization of these massive datasets raises issues of security, confidentiality, and data ownership.