Don’t Let AI Skip Brain Day

You open ChatGPT to write an email you could probably have handled yourself. A few minutes later, it summarizes the article you meant to read, suggesting ideas for your presentation and helping you decide which one is best. By the afternoon, you might even be asking it to phrase the awkward text you have been putting off.

None of this is inherently bad. AI is useful because it saves time and effort, often on tasks that are tedious, repetitive or genuinely difficult. The concern begins when “help me with this” quietly turns into “think through this for me.”

Researchers sometimes call this “cognitive offloading”, which simply means shifting mental work to an external tool. We have always done it: calculators handle arithmetic, GPS remembers the route, and search engines retrieve information we no longer keep in our heads. Generative AI goes further because it can take on several kinds of thinking at once, including writing, reasoning, brainstorming, research and decision-making.

The early research is complicated, which is precisely why the conversation should not be reduced to “AI makes us stupid.” In one much-discussed MIT Media Lab study, people writing essays with an LLM showed lower measures of neural engagement and had more trouble recalling their own work than those writing without digital assistance. Other research points in a more encouraging direction. A 2026 randomized experiment found that students given access to generative AI performed better on tests of what they had learned, with some gains still visible a week later. The interesting part was how they used it: people who asked AI to explain concepts and help them understand tended to benefit more than those who mainly used it to generate text for them.

That may be the more useful way to think about the whole debate. The question is not simply whether you use AI, but whether you use it to augment your thinking or automate it away. Here are five ways to keep yourself on the right side of that line.

Think first, prompt second

Start with timing. There is a meaningful difference between bringing in AI after you have engaged with a problem and opening it before you have even tried.

Before typing the prompt, give yourself a few minutes to form an opinion, sketch an answer or work out the first step. If you are writing, get something onto the page. If you are making a decision, work out where you are leaning. If you are trying to understand an idea, explain it to yourself first, even badly.

That initial struggle can feel inefficient when AI can produce something polished in seconds, but it tells you what you know, what you do not know and where you are stuck. Skip that stage every time and you risk becoming very efficient at producing finished work without getting much better at producing the thinking behind it.

A useful rule remains: brain first, bot second.

Use AI to challenge your thinking

Write your argument, then ask AI for the strongest objection. If you are studying, ask for a hint rather than the solution. If you are preparing a presentation, ask which slide is least convincing rather than asking it to build the whole deck. You can ask it to play the sceptical boss, difficult professor or colleague who keeps questioning your assumptions.

Used this way, AI gives you something to respond to rather than something to copy. The emerging evidence around learning supports this distinction: using AI for explanation and understanding may produce different outcomes from simply using it to generate the finished output.

Keep judgement in the process

That makes verification more than a fact-checking chore; it is part of the thinking you should retain. For anything consequential, check where the claim came from, whether the statistic is current and whether credible evidence points in another direction. As generating information becomes easier, deciding what deserves to be believed becomes more valuable.

Protect the skills you want to build

If you are trying to become a better writer, outsourcing the first draft every time defeats part of the purpose. If you are learning financial analysis, having AI perform all the analysis may give you a beautiful answer without much practice analysing. If you want to become better at generating ideas, ask AI to expand or challenge your ideas rather than supplying all of them.

Automate the peripheral work where it makes sense. Let AI format, organise, transcribe, compare or help you search. Be more cautious about handing over the specific mental skill you hope to get better at.

Notice when AI becomes a habit

None of those behaviours automatically means you are using AI badly. The better test is whether the tool is expanding what you can do or gradually becoming a prerequisite for things you once did comfortably yourself.

If AI has become your starting point for almost everything, introduce a little friction. Make the rough first attempt, guess before checking or spend a few extra minutes sitting with a difficult problem. Sometimes the slower route still has value because the process, not just the answer, is what you are trying to keep.

AI will inevitably take more work off our hands. The useful question is which work we are happy to hand over, and which parts of thinking we still want to be good at.

Fardeen Ameen is the the former Head Of Gaming Business at Grameenphone; presently pursuing an MBA at Bayes Business School in London by day and holding onto being an unapologetic gaming, tech & sports fanatic by night.