Paul W. Bennett: AI isn't revolutionizing learning. It's mimicking original thought

National Post

31-07-2025

Generative AI platforms such as ChatGPT have entered classrooms, universities, and homework routines with astonishing speed and little attention to the long-term consequences. A recent Canadian news report, aired on CBC's The National, also revealed that teachers have mostly been left to fend for themselves.
Article content
The current infatuation with AI is part of a recurrent pattern, but the latest educational fad is far more fundamental in its impact on teaching and learning in classrooms. It's time to ask: Are these tools eating away at our brain power and leading schools astray?
Article content
Article content
Article content
Technology evangelists and educators espousing '21st century learning' tout its ability to save time, individualize instruction, and increase access to information. But little has been done to assess its effects on students' ability to think independently, write clearly, and engage with knowledge deeply.
Article content
Article content
What's encouraging is the fact that leading cognitive scientists, evidence-based researchers, and experienced frontline teachers are beginning to right the balance.
Article content
There is mounting evidence that the emergence of ChatGPT and similar AI tools are short-circuiting deeper learning, eroding critical thinking capacities, and undermining the teaching of writing. Our brains, it turns out, need knowledge to function at their best.
Article content
Generative AI is proving to be a large learning model which encourages passivity in learners. Leading cognitive scientist, Barbara Oakley, an American expert on learning how to learn, warns that ' mental effort is essential to build real understanding.' Meaningful learning, according to Oakley, is built through deliberate practice, cognitive struggle, and retrieval of knowledge — all processes undermined when students delegate intellectual labour to AI tools.
Article content
Article content
Bypassing the productive discomfort associated with writing and problem-solving, students risk becoming consumers of content rather than producers of thought. The process of wrestling with an argument, organizing one's thoughts, and finding the right words is foundational to critical thinking.
Article content
Article content
Generative AI, however, short-circuits this developmental trajectory by offering polished outputs without much heavy lifting. If students become accustomed to outsourcing the most demanding aspects of thinking and writing, known as cognitive offloading, they lose the capacity to do it themselves.
Article content
American education commentator, Natalie Wexler, author of The Knowledge Gap, sees AI as the latest educational trend that emphasizes skills over content, inhibiting our capacity to grasp and understand knowledge in context. True critical thinking, she argues, cannot be taught in isolation from a deep base of knowledge. In her view, students need a well-stocked mental library of facts, concepts, and contexts to think critically and write effectively. Generative AI, by providing surface-level responses to prompts, may reinforce the illusion that knowledge is readily available and easily synthesized, even when it lacks depth or coherence. Students may come to view knowledge acquisition as unnecessary, assuming that AI can fill in the gaps. This undermines both the cognitive effort required to develop coherent explanations and the long-term retention that underpins higher-order thinking and genuine problem-solving.