THE USEFUL BIT

AI Is Making It Too Easy to Avoid Thinking (And My Brain Is Thrilled)

I asked AI to help me write a joke about people who use AI instead of thinking. It returned five versions, explained why each one was funny, and suggested an A/B test. By the time I finished reading, the punchline had already applied for a job in marketing.

That is the strange convenience of AI: it can draft the email, plan dinner, summarize a long document, brainstorm a business idea and explain a joke. Soon, the hardest part of having a thought may be deciding which chatbot should have it for us.

At this point, some of us do not think through a problem. We create a prompt about thinking through it, then ask AI to improve the prompt. Our brains' new job is approving the answer with the confidence of a manager who has not opened the attachment.

The joke has a serious little footnote

Researchers use the term cognitive offloading for using tools or actions to reduce the mental work a task requires. Writing a reminder instead of memorizing an appointment is a familiar example. Offloading is not automatically bad; tools can free us to focus on something more important.

The question is what happens when we offload the very part we are meant to practise: remembering, reasoning, checking or forming our own view.

One 2025 randomized study of undergraduates learning AI compared ChatGPT-assisted study with traditional study methods. On a surprise knowledge-retention test 45 days later, the AI-assisted group scored 57.5% correct, compared with 68.5% for the traditional-study group. That is worth paying attention to, but it is one study in a specific learning setting—not proof that every AI user is losing brain cells by the minute.

A 2026 study points to a more complicated picture. In a task identifying logical fallacies, participants using ChatGPT got more answers right and reported lower mental effort. Its authors noted that AI does not uniformly improve or weaken critical thinking; the way people use it—as a replacement for their reasoning or as support for it—matters.

In other words, the research does not say “never use AI.” It says the answer depends on the task and on whether you are still mentally in the room.

How to use AI without putting your brain on airplane mode

  • Try first, then ask. Write down your own answer or idea before asking AI. Even a messy first attempt gives you something to compare.
  • Request coaching, not just a finished answer. Ask for hints, counterarguments or questions that expose gaps in your reasoning.
  • Make it show its work. Ask what assumptions it made, then verify important facts and sources independently.
  • Explain the result in your own words. If you cannot explain the answer without reopening the chat, you may have borrowed the conclusion without learning much.

AI is brilliant for reducing repetitive work. It is less brilliant as the final judge of what matters, what is true or what you believe. Keep your taste, curiosity and healthy skepticism in the loop.

A good rule: let AI carry the groceries, but do not ask it to chew your dinner and tell you how delicious it was. You still have to taste the answer.

Sources

✳
Keep your curiosity going.

Good information should help you ask better questions, not stop asking them.