The integration of artificial intelligence (AI) into education has raised many questions about how best to use this powerful technology. On October 13, the fifth edition of Métamorphose(s) de l’Éducation, organized by Onepoint, will focus on the theme "The Great Awakening: AI, Tech, Human Wealth." The event will examine the challenges and opportunities of using AI in schools and will bring together experts from various fields, including researchers, teachers, entrepreneurs, and philosophers, to discuss its future role in education. Sociologist Pierre-Michel Menger has warned that relying too much on AI might lead to a misunderstanding of what learning truly is. He argues that knowledge is not just about getting the right answer but also about the process of searching, comparing, and making mistakes. Muriel Touaty, who works in education, research, and innovation at Onepoint, describes AI as an "exoskeleton of thought"—a tool that can support human thinking but could also weaken it if people depend too much on machines for reasoning. At the event, there will be discussions about how AI is changing the way students are taught. Prominent participants include Jean-Michel Blanquer, Frédérique Vidal, and Olivier Ginez. One panel will debate the question "Educate or train: what school do we really want?" Philosopher Charles Pépin will discuss how to "train a human in the era of AI," while philosopher Éric Sadin will ask a more radical question: "Should we unplug AI?" These discussions will explore the broader implications of AI in shaping the future of education. AI systems, such as ChatGPT, Claude, Perplexity, Gemini, and Mistral Vibe, have the potential to help students and teachers by offering resources, explanations, translations, and support in creating texts and presentations. However, they also raise concerns about the loss of critical thinking skills and the risk of reducing teachers to mere assistants. If students rely too heavily on AI for cognitive tasks, they may become less able to think independently. Additionally, the use of AI could change the social aspects of education by making learning more individualized, which could increase inequalities among students. One proposed solution is "educational bilingualism," an approach that divides teaching into two tracks: one that avoids AI entirely and one that uses it extensively. This method aims to preserve traditional learning methods while also developing digital skills. By comparing the outcomes of both tracks, educators can better understand the benefits and drawbacks of AI in education. This approach also offers a way to test AI's effects and protect against risks like technical failures, dehumanization, and long-term impacts on learning. The debate about whether AI is helpful or harmful in education is still ongoing. Some educators see AI as a powerful tool that can enhance learning, while others are concerned about its potential negative effects. Educational bilingualism offers a way to test both approaches in real classrooms, avoiding the limitations of hybrid models that only use AI in small amounts. This method allows for a more thorough evaluation of AI's role in shaping the future of education.