Skip to Content, Navigation, or Footer.
Support independent student journalism. Support independent student journalism. Support independent student journalism.
The Dartmouth
March 4, 2026
The Dartmouth

Reflection: Detoxing from AI

One writer walks through a work week spent away from a technology he — and countless others — has come to rely on.

student using computer

In the last two years, I don’t think I’ve gone a single day without hearing the term “AI.” Every time I open Instagram or YouTube or even have a conversation with someone, artificial intelligence is bound to come up. And for good reason. Personally, AI has been part of my routine for a long time — especially as a computer science and math major. I’ve used it for years to help with coding, problem sets  and helping me research, going as far back as when Gemini was still called Bard. But I’ve become tired of AI because I feel like it’s inhibiting our learning.

I wanted to find out if I could last five days in a Dartmouth term without AI, so I quit cold turkey. 

Throughout the week, I learned more about my own capabilities, and that while AI may be a useful tool, it has a tendency to hijack original thinking.


Monday

Usually, I spend my mornings preparing for my class, MATH 022: Linear Algebra with Applications. I start with the textbook and use Google Gemini to clarify what I don’t understand, but this time I had to rely on the text and hunt through YouTube for the right explanations. I was worried this would take significantly longer than my usual process — videos have a tendency to re-explain things you already understand. 

Luckily, that wasn’t the case. I found a linear algebra course broken into short, five-minute videos, each focused on a specific subtopic. I could jump around freely, and my understanding of the material didn’t noticeably suffer.

I was surprised by how focused I felt. Normally, while AI is processing my question, I’m scrolling through Instagram or checking notifications. Without AI in the picture, that gap never opened. I wasn’t switching between tabs on my laptop anymore, which felt a lot more productive.


Tuesday

I woke up on day two afraid. The first draft of my final paper for Writing 5 was due, and it was a substantial piece of work.

I generally don’t like using AI to write for me, but I hadn’t fully appreciated how much I rely on it in the early stages of the writing process. Normally, I use it to generate specific search terms I can plug into databases like Google Scholar or JSTOR. Without that scaffolding, I felt incredibly lost. The topic I was researching didn’t have information compiled in a centralized place, and I didn’t have a clear alternative strategy. I ended up just having to grind through it — spending far more time on the essay than I had planned. Not being able to use AI also meant I couldn’t rely on tools like Grammarly to restructure my sentences.

Relying solely on my writing skills — along with a lot of Googling and Thesaurus websites — gave me an important realization. My writing hasn’t improved since the pre-ChatGPT era. The foundations I’ve been working with were built then, but even with increased exposure to writing now, the introduction of AI has stunted my growth as a writer. Without that effort of struggling to phrase sentences, fix awkward structure, or work through unclear ideas on my own, the pressure to meaningfully get better over time substantially fell.

There’s also something ironic about the whole arrangement: AI was supposed to help me get rid of the menial work and focus on the thinking, but the more I rely on AI, the more I become the one simply prompting, while the machine does the actual intellectual work.


Wednesday

I woke up looking forward to this day of work. I had a lot of homework to get through, and I was curious to see how I’d hold up.

For math, I usually solve problems and have Gemini check my work, but without it I had to verify things manually and leaned more on my friends in the same class, which was slower but probably more useful. Talking through problems forced me to articulate my reasoning, which exposed gaps I would have otherwise missed, and conversations often led to better ideas than working alone.

I also spent time on a personal coding project. I expected this to be the area where my loss of skill would finally embarrass me. Tools like Claude Code usually handle a lot of the repetitive, mechanical work in my projects, and I assumed that without them, the gaps in my own knowledge would become obvious. That didn’t happen. The code I wrote held up, the logic came back to me naturally, and for debugging I found myself on StackOverflow — a problem-solving forum website that felt almost nostalgic to use.

The most surprising moment of the day had nothing to do with any of my classes. That evening, my friends and I were sitting in the corner of Foco late night trying to figure out how many Lego bricks could fit in my dorm room. I pulled out my phone to look up the dimensions of a standard Lego brick, and without even thinking about it, I opened ChatGPT instead of Google.

I caught myself and paused. When did that happen? At some point, without noticing, ChatGPT had simply replaced Google as my default way of looking things up.


Thursday

By Thursday, my brain adjusted to my workflow and the temptation to open a new tab and ask Claude something practically disappeared.

The more I sat with my work on my own on Thursday, the more I became aware of how rare it is to do so for myself and my peers. My peers were still using AI freely, moving faster and producing things in seconds that were taking me considerably longer. I felt a gap in my output. In an environment where AI has become so embedded in how people study and produce, opting out, even temporarily, carries a cost. It’s hard not to feel like you’re falling behind.


Friday

The last day came and went more quietly than I expected. Just a Friday wrapping up a relatively stress-free week.

I’d been looking forward to having AI back, to be honest. As someone who has nearly entirely relied on self-studying my entire life, AI has helped improve my learning process by leaps and bounds. It has helped me understand difficult material faster, structure my thinking and vastly improve the quality of my work. 


Final Thoughts

After this week, I will never approach AI in the same way. In some ways, I think AI is poisoning the very concepts it was helping me learn so well. I realized that AI has been subtly ruining my cognitive abilities — and I wasn’t even realising it. In a way, AI feels like a drug. 

Every time you open up Chat, or Gemini, or Claude, and ask it a question, it usually gives you the answer you were looking for. Somewhere in that exchange, though, a small amount of critical thinking,  understanding, slips away without the user noticing. In the chase for efficiency and convenience, I was losing the one thing that can differentiate me from an AI: my cognition and creativity.

AI is almost too good at its job. It is so finely tuned to give you exactly what you’re looking for that it leaves very little room for the cognitive effort that makes you better over time. But more than that — you can’t really choose to avoid it. Everything from a Google search to my headphones app now has some kind of AI feature baked in. Truly opting out is becoming harder by the day.

I’ll generally keep using AI. I don’t think I’d be able to keep up without it. I’ll use it to code projects, to organize my calendar, to learn new concepts, as a repository of information. Though AI as a tool will re-integrate itself into my life, I hope that my new awareness will allow me to step back — if not abandon completely — a technology that has slowly dominated our lives.