I spent two years learning about artificial intelligence-based tools. I read the papers, tested out the models, attended the workshops, built prompts, compared outputs, and watched each new release with the excitement of someone witnessing the future unfold in real time. Like many people working close to technology, I believed AI would become the defining aspect for our generation.

Then, I spent one year unlearning AI. I realised how astonishing these systems had become. I also realised that the greatest threat AI posed was not to my career, but to my mind. This latter realisation came when I uninstalled ChatGPT from my phone.

My friends assumed I was making some grand statement against technology. But I was simply trying to rescue a part of myself that had begun disappearing. The habit of turning to an AI tool starts innocently enough. You want to Google something, so you ask ChatGPT. You struggle to begin writing an email, so you ask ChatGPT to write it for you. You cannot quite articulate a thought, so you ask ChatGPT for help. Eventually, the question stops being “what do I think?” and becomes “what does AI think I should think?”

I had spent years believing AI was merely another tool. But tools usually help us do things better. Generative AI, on the other hand, offers to do the thinking before we have even attempted to do it ourselves. And most of us avoid admitting how dangerous this bargain is. For centuries, good education has been built around a simple principle: struggling with an idea so you can make it your own. Every confused student, rewritten paragraph, abandoned draft, and wrong answer contributes to the better understanding of things. AI offers us a world without this essential friction and, by extension, without the process of thinking. Psychologists call this cognitive offloading, which means to shift mental effort onto external tools (in this case, generative AI).

As one would expect, after deleting ChatGPT, my life became slower. The first few days, I absentmindedly swiped through my phone looking for the app. But eventually, I began spending longer browsing through the internet. I also spent a considerable amount of time talking to people. I found myself reading instead of seeking summaries. Tasks which once took an hour now consumed entire evenings. For a while, I feared I had become less intelligent.

But soon I started remembering things again. My ideas stayed with me because I had earned them. My arguments became sharper because I had built them myself. I stopped confusing instant access to information with genuine understanding. Most importantly, I started writing again. Not just generating text; actually writing.

Writing has never been about arranging words into grammatically correct sentences. It is instead the slow and often frustrating process of discovering what one actually believes. Every first draft is a conversation between conflicting ideas; every revision is evidence that one’s mind is growing. Generative AI skips all of that and delivers polished conclusions without any intellectual struggle. Some term this efficiency, but it is actually intellectual outsourcing. Somewhere along the way, we began celebrating people for producing more instead of thinking better. Students submit essays they never wrestled with. Marketers publish campaigns they never conceived. Professionals proudly announce that AI writes their emails, prepares their presentations, and even generates their reports.

We have become curators of machine-generated thoughts instead of authors of our own. The most worrying aspect is that this dependence feels rational in isolation. Why spend three hours writing when AI can do it for you in 30 seconds? Why read an entire book when a detailed summary can be generated? Why struggle through uncertainty when certainty is so accessible? But uncertainty is where thinking happens. That very uncertainty and the struggle that comes with it are doing something an algorithm can’t: they are helping us develop our intellect.

Every meaningful belief I hold today was shaped by confusion that was followed by clarity. Every article I am proud of began as an incoherent mess. Every worthwhile conversation I’ve had forced me to reconsider assumptions I wasn’t even aware I harboured. None of these could have been “prompted” into existence.

People often ask whether I have tried Claude, Gemini, or whichever new AI tool currently dominates the headlines. The truth is that I have not. But the question should really be: what kind of person do I become if I allow a machine to perform the very activities that shape who I am? Thinking is not merely a means to an answer. It is how we construct our own identity.

I increasingly believe history will remember this era not as the moment machines became capable of writing, but when humans voluntarily stopped trying. That, to me, is infinitely more tragic. We are told that AI will free us to focus on higher-order thinking. But everywhere I look, the opposite is happening. After all, AI tools encourage us to surrender the very habits from which higher-order thinking emerges.

Perhaps the saddest consequence is that we no longer trust ourselves. Faced with an empty page, many people's first instinct is not to think but to ask a chatbot about what they ought to say. We have begun borrowing confidence from machines because we no longer believe our own imperfect thoughts are enough. I now refuse to live that way; to become a second-hand thinker in my own life. I would rather write a flawed article than publish a “perfect” one that I cannot really claim ownership of. I would rather forget, search, revise, and begin again than slowly surrender the most human thing I possess: my ability to think.

-The Daily Star (Bangladesh)/ANN