We like to believe that thinking is an essential human trait, something irreducible and entirely our own. Yet with the rise of artificial intelligence, that assumption is beginning to feel less certain and more negotiable. The concern itself is not new. When calculators became common, people worried about losing arithmetic skills. When search engines arrived, some feared the erosion of memory. When autocorrect and predictive text became standard, some worried that our ability to spell and construct sentences would weaken. The brain, it seems, is efficient to a fault. It does not hoard what it can reliably retrieve. AI may simply be the latest and most capable extension of this tendency.

A 2011 study by Betsy Sparrow and her colleagues at Columbia University showed that when people expect information to be available externally, they are less likely to encode it internally. This phenomenon, often called the “Google effect,” reveals how memory is not just storage. It is strategy. We remember what we believe we will need, and we forget what we believe the world will remember for us. AI intensifies this dynamic. It does not just store information. It interprets, organizes, and even anticipates. In doing so, it invites us to step back from processes that once demanded direct cognitive effort. 

But the shift is not only about memory. It is about the experience of thinking itself. Psychologists such as Daniel Kahneman have long distinguished between fast, intuitive thinking and slow, deliberative reasoning, arguing in Thinking, Fast and Slow that the latter requires effort precisely because it resists immediacy. AI, by contrast, is built for speed. A study by researchers at Carnegie Mellon University and Microsoft found that while AI tools can improve productivity, they can also reduce the amount of critical evaluation users apply to tasks, with participants less likely to question outputs or explore alternative approaches. In educational contexts, this raises a subtle but significant concern: if students arrive at correct answers without traversing the cognitive path that produces them, what exactly have they learned?

This distinction places responsibility not on the technology but on its use. Educational theorists increasingly emphasize the importance of “desirable difficulty,” a term popularized by Robert Bjork, which refers to challenges that slow learning in the short term but deepen it in the long term. AI, when used uncritically, removes these difficulties. Yet when used deliberately, it can also create new ones, prompting comparison and critique. The question, then, is not whether AI diminishes thinking, but whether it alters the balance between effort and ease in ways we do not fully recognize. In a world where answers are immediate, the discipline to ask better questions may become the more valuable and more endangered skill.

In the end, the question is not whether AI will think for us, but whether we will continue to think for ourselves. Tools have always changed the way we use our minds, but rarely have they come this close to replacing the process itself. If thinking becomes optional, it may also become rare. And what is rare often becomes valuable, but not always preserved.