For years, those of us in educational technology promised that digital tools would transform schools. Well, the transformation is finally here, but it did not arrive as a gentle evolution. It hit us like a shockwave. We are now facing what I call the "Ghostwriter Paradox". In the old days, a ghostwriter was a luxury, someone paid to write a book while another person took the credit. Today, Generative AI has become the world’s cheapest, most available ghostwriter. Here is the problem: When a student asks a chatbot to author an essay, the student’s name goes on the paper, but the actual thinking happens inside the machine. Writing isn't about putting words on a page; it is a struggle that forces us to organize our thoughts. When we outsource that struggle to an algorithm, students lose the chance to build their own mental structures. They risk becoming mentally lazy, acting as passive verifiers instead of active thinkers. So, what do we do? First, we need to stop talking about "AI Literacy" and start talking about "AI Fluency". Literacy is just knowing how to use the tool, how to type a prompt or log in. That is the easy part. Fluency is deeper. It means understanding that the AI does not actually "know" anything. It is just a "stochastic parrot", a fancy way of saying it guesses the next word based on statistics. If a student understands that the bot is guessing rather than thinking, they are less likely to trust it blindly. Second, we must be honest about our exams. A good rule of thumb for 2026 is this: If a machine can pass your exam, your exam is broken. We cannot keep grading take-home summaries. We are just grading the AI. We need to focus on the process, not the final product. This means bringing back old-school methods like oral defenses or having students write in class. Ask the student to defend the idea. If they cannot explain it, they didn’t write it. To make this work in real schools, I recommend we look to the practical '4P Framework' from the Department of Education in Ireland (2024). It offers a clear roadmap: • Protection: This is the baseline. We must never feed sensitive student data into public AI models. We also need to protect students from getting too emotionally attached to chatbots that simulate empathy but have no conscience. • Policy: We cannot just ban these tools. We need clear rules that keep the "Human in the Loop". Policies should clarify the difference between using AI as a scaffold (like a tutor) and using it as a substitute (like a ghostwriter). • Pedagogy: We need to teach about AI, not just with it. Open the "black box. " Show students how the models work and let them try to trick or "break" the AI to see its biases. • Practice: Teachers cannot do this alone. We need to stop doing one-off webinars and start creating communities where teachers can experiment, fail safely, and learn together. The answer to AI passivity is human activity. We need to move away from valuing how fast a student can get an answer and start valuing whether they know if the answer is true. The AI is a great statistician and a wonderful mimic. But it is not a thinker. That role still belongs to our students, and it is our job to make sure they keep it. Reference Department of Education Ireland. (2024). Guidance on artificial intelligence in schools. Government of Ireland. https: //www. accs. ie/s/DEY-GuidanceₒnArtificialIntelligenceᵢnSchools₂025. pdf
Stamatios Papadakis (Tue,) studied this question.