AI and Critical Thinking: What Happens When Answers Become Effortless?

Artificial intelligence is becoming part of learning so quietly that we may not yet appreciate what is changing.

Students are asking AI to explain concepts, solve problems, summarize information, generate ideas, and complete assignments. What was once considered an emerging technology is quickly becoming an everyday learning companion. According to a 2025 Higher Education Policy Institute (HEPI) and Kortext survey, 92% of university students reported using generative AI, up from 66% just one year earlier.

Much of the conversation around AI and critical thinking asks whether AI is good or bad for education. today asks whether AI is good or bad for education. Perhaps there is another question worth observing.

Is AI changing education, or is it changing learning itself?

There are good reasons to be optimistic. Organizations such as UNESCO and the OECD have highlighted AI's potential to personalize learning, improve accessibility, provide immediate feedback, and support learners at their own pace. During our first EFC Perspectives conversation, a similar thought emerged. AI has the potential to make learning more individualized than ever before. For children who already enjoy exploring ideas, asking questions, and thinking independently, AI may become an extraordinary partner in learning.

But another thought stayed with me.

Over the past few months, we've explored the difference between completing work and developing understanding. AI introduces a new dimension to that same conversation.

If answers become easier to obtain, what happens to the thinking that once sat between the question and the answer?

Recent research from Microsoft Research suggests that the relationship between AI and critical thinking is more nuanced than many expected. Greater confidence in generative AI systems may reduce the amount of critical engagement people apply to AI-generated responses. The technology continues to improve at producing answers. Whether it also strengthens the thinking behind those answers is a much more interesting question.

It becomes even more interesting when viewed alongside what we already know about learning.

According to the OECD's PISA 2022 assessment, only 66% of U.S. students reached baseline mathematical proficiency, the level at which students begin demonstrating the ability to apply mathematics in simple unfamiliar situations. Long before AI entered classrooms, many learners were already finding it difficult to transfer knowledge beyond familiar examples and practiced patterns.

AI did not create that challenge.

But it now enters a learning environment where those foundations already exist.

For some children, AI may deepen understanding.

For others, it may make it easier to move forward without fully developing it.

And perhaps this is where my curiosity continues to grow.

If AI makes information and answers increasingly abundant, perhaps the nature of learning begins to shift as well. The challenge may no longer be finding an answer, but understanding it. The ability to reason through unfamiliar situations, question assumptions, connect ideas across disciplines, and exercise sound judgment may gradually become more valuable than simply arriving at the correct response.

If that is true, another question naturally follows.

For generations, we have measured learning by asking children to produce answers. But if answers can increasingly be generated with AI,

how might we begin observing understanding in ways that cannot simply be outsourced?

Perhaps the future of AI and critical thinking is not about limiting AI. It may be about helping children develop the kind of thinking that allows AI to remain a tool rather than a substitute for understanding.

Perhaps the question is no longer whether AI belongs in education.

AI is already here.

Perhaps the more important question is whether we are helping children develop the kind of understanding that allows AI to remain a tool, instead of quietly becoming the thinking itself.

I don't think we know the answer yet.

But it feels like a question worth observing.

And perhaps one worth observing together.