AI Can Make You Smarter: How to Use It Without Losing Skills

Using AI can feel like having an extra brain on demand.

At first, the benefits are obvious. You can write emails faster, solve problems more quickly, organize information, and get through tedious work in a fraction of the time. For many people, that boost can make them feel more productive, confident, and even smarter.

But there can be a downside.

After relying heavily on AI, some people notice that their memory isn’t as sharp, their writing feels weaker, or they struggle more with problem-solving and decision-making. The uncomfortable part is that these changes can happen gradually, without you realizing how much thinking you’ve handed over to a chatbot.

Susette Brooks, a marketing director in Philadelphia with an MFA in nonfiction writing, experienced this firsthand. She initially resisted AI tools but eventually started using them to write some of her emails. While the technology reduced her workload, she later noticed a decline in her own writing and thinking abilities. She also found it harder to recall information and connect ideas.

The issue has a name: cognitive offloading.

What Is Cognitive Offloading?

Cognitive offloading happens when we use an external tool to reduce the mental effort required to complete a task.

You’ve probably done it without thinking about it. Using a calculator instead of doing arithmetic in your head is cognitive offloading. So is putting an appointment on your calendar rather than relying on your memory.

AI takes this idea much further.

Tools such as ChatGPT and Claude can calculate numbers, remember information, summarize documents, write emails, create presentations, solve problems, and even produce code. In other words, we’re no longer just outsourcing small pieces of mental work. We’re increasingly outsourcing entire cognitive processes.

That isn’t automatically a bad thing.

Cognitive offloading is one of the reasons humans can accomplish so much. External tools allow us to overcome the limitations of our memory and attention.

The problem starts when we consistently outsource tasks that we actually need to practice ourselves.

That’s where de-skilling comes in.

AI Can Boost Productivity While Weakening Certain Skills

De-skilling describes the gradual loss of competency when people stop practicing skills because technology performs those tasks for them.

With AI, that could mean writing less, doing less independent analysis, remembering fewer facts, or spending less time working through difficult problems.

The effect is particularly important for people who are still developing those skills.

Key takeaways

• AI can make you more productive while quietly reducing your ability to work independently.
• Offloading repetitive or tedious tasks is generally less concerning than outsourcing important thinking and skill-building.
• The answer isn’t necessarily to use less AI. It’s to be more deliberate about which tasks you give to AI.

Young People Are Already Worried About AI and Thinking Skills

The concern isn’t limited to researchers.

A Gallup survey conducted with researchers from the University of Pennsylvania found that 68% of Gen Z adults in the United States worry that outsourcing too many cognitive tasks to AI could deprive them of important opportunities to build skills.

A similar proportion said they were concerned that AI could discourage critical thinking and social learning.

A separate Pew Research Center study published in June found that 37% of U.S. adults under 30 believe AI will have a negative effect on them personally.

And yet, more than half of Gen Z Americans use AI every week.

That’s the central contradiction: people can recognize the risks while still finding AI far too convenient to ignore.

Why Our Brains Prefer the Easy Option

Daniel Willingham, a psychology professor at the University of Virginia, argues that humans are naturally inclined to avoid unnecessary cognitive effort.

Thinking is demanding. It takes energy, and there’s no guarantee that all that effort will lead to the correct answer.

Memory is different. If you’ve already learned that something worked before, recalling it is usually much easier than working everything out from scratch.

As Willingham puts it, our minds are essentially designed to minimize the amount of thinking we have to do.

AI fits perfectly into that tendency.

Why struggle through a difficult problem when you can ask a chatbot for the answer in seconds?

That’s exactly why using AI thoughtfully matters.

Too Much AI Use Can Affect How Well You Learn

Derek Salyers, a content creator in Ohio, started using AI because he wanted to accomplish more in less time.

Eventually, he noticed two problems.

First, the quality of his work declined. He described some of what he produced as “AI slop.” More importantly, he felt his own thinking was deteriorating.

When AI makes decisions for you repeatedly, you can stop pausing to make those decisions yourself. And if you never actively process the information, you’re less likely to remember it later.

This isn’t necessarily a failure of memory. Sometimes, the information simply never received enough attention to become something you could recall.

Research involving students points in a similar direction.

One study of nearly 1,000 high school students found that students using ChatGPT scored 48% higher on a math test than students who didn’t have access to the tool. But when the students were tested without ChatGPT, their scores were 17% lower than those of the control group.

Researchers at the University of California, Irvine also examined longer-term data from online tests beginning in 2015. Students spent less time on text-based questions and performed better after ChatGPT became available. But when they later completed tests offline without AI assistance, their performance was worse.

The pattern suggests that AI can produce immediate performance gains while potentially reducing the amount of independent cognitive work people perform.

Not Every Mental Skill Is Equally Vulnerable

There is an important distinction here.

Basic cognitive abilities such as working memory and selective attention are relatively resilient. But learned skills are different.

Arithmetic, writing, coding, and other complex abilities require practice. You first need to acquire the skill, and then you need to keep using it if you want to maintain it.

Research from Evan Risko, a psychology professor at the University of Waterloo, suggests that cognitive offloading can be useful when it extends our limited memory and attention. Setting reminders, for example, frees us from having to remember every appointment ourselves.

The risk comes when we outsource mental processes that are important to maintain.

A paper co-authored by Risko and colleagues, titled “Is AI making us stupid?”, argues that AI can interfere with both learning and retaining certain skills.

There’s another problem: AI can make people appear more capable than they actually are.

Brooke Macnamara, a psychology professor at Purdue University and co-author of the paper, notes that people may not realize their independent skills are declining because their performance remains strong while AI is available.

If the chatbot is always there to help, you may never discover how well you can perform without it.

The Best AI Strategy? Be Selective

The most important question to ask before using AI is surprisingly simple:

Do I actually need AI for this task?

Anjali Singh, a human-AI interaction researcher at the University of Texas at Austin, argues that simply having an AI tool available doesn’t mean it should be used for everything.

Writing is a good example.

If AI writes your emails, reports, essays, or articles from scratch, you may save time. But you’re also giving up the thinking involved in organizing your ideas and expressing them clearly.

And if you don’t practice writing, your writing skills have fewer opportunities to improve.

That’s how de-skilling can develop over time.

Use AI for Skills You Already Have

One of the more counterintuitive recommendations from researchers is to use AI most heavily on tasks you’re already good at.

A working paper from MIT economist David Autor and colleagues examined how 133 patent lawyers performed with and without AI assistance. Several researchers involved in the work are employed by Google, which funded the study.

The results were interesting.

All of the lawyers benefited from AI while they were using it. But after the tool was removed, younger lawyers lost those gains. More experienced lawyers, meanwhile, performed better afterward than comparable lawyers who had never used the AI tool.

The lesson isn’t that AI is bad for inexperienced workers.

It’s that experience matters when deciding how AI should be used.

Another experiment involving more than 200 college students at Middlebury College found a similar distinction. Students learned an unfamiliar subject and wrote analytical essays, with some using a standard AI tool.

A week later, they completed a similar test without AI.

The students who had used AI improved overall, but researchers identified two different approaches. Some students used AI to augment their own thinking, while others used it to automate the work.

The augmentation group performed best.

That’s an important distinction: AI can help you think, or it can think instead of you.

The first approach is far more likely to strengthen your abilities.

Tell AI to Teach You Instead of Doing Everything for You

There’s another way to change your relationship with AI: stop treating the chatbot like an answer machine.

If you’re trying to learn something complicated, asking AI for a massive explanation can leave you with even more information and less understanding.

Older methods of learning, such as reading books or researching through Google, were slower. That slower process created natural opportunities to stop, think, compare information, and form your own conclusions.

AI removes much of that friction.

So bring some of it back.

If you’re using ChatGPT, Claude, Gemini, or another AI assistant to learn, tell it that your goal is to build the skill rather than outsource it.

Ask the AI to:

• Quiz you instead of immediately giving you the answer.
• Ask follow-up questions.
• Give hints when you’re stuck.
• Explain why an answer is correct.
• Challenge your assumptions.
• Provide practice problems.
• Give feedback on your work.
• Help you identify gaps in your understanding.

ChatGPT’s Study Mode, Claude for Education, and Gemini for Students are examples of AI tools designed around this more interactive approach.

The goal is to make the AI behave more like a tutor than an answer generator.

Slow Down and Keep Some Friction

One of AI’s biggest attractions is speed.

Tasks that once took hours can sometimes be completed in minutes. Coding is a good example. With AI-assisted or “vibe” coding, you can create an application much faster than you could by writing every line manually.

That’s useful.

But faster isn’t always better for learning.

Joei Chen, a former technology company brand director, described how heavy workloads can make people reach for AI without stopping to think. When there isn’t enough time to process a problem yourself, it’s easy to simply ask the chatbot what it thinks.

The danger is that AI can gradually become the final decision-maker.

A recent Wharton School study also found that producing more code doesn’t necessarily translate into more completed software, while higher productivity doesn’t automatically make workers happier.

The solution isn’t to reject AI.

It’s to introduce some friction back into the process.

Instead of asking AI to write something and accepting the first result, ask for several options. Read them all. Compare them. Decide what works and what doesn’t.

That small amount of additional effort forces you to exercise judgment.

Make AI a Collaborator, Not a Replacement

A useful rule is to make your AI interactions collaborative rather than transactional.

Don’t just ask:

“Write this for me.”

Try:

“Give me five possible approaches and explain the strengths and weaknesses of each.”

Don’t ask only for the answer. Ask for feedback, examples, explanations, hints, and practice questions.

Advait Sarkar, an AI researcher at the University of Cambridge, argues that simply comparing different options can stimulate important critical-thinking processes. It can also improve your judgment, awareness of what you don’t know, and ability to recall information.

This approach also encourages metacognition, or thinking about your own thinking.

That idea came up repeatedly among the experts discussing how people can use AI without becoming overly dependent on it.

The more AI does everything automatically, the less opportunity you have to examine how you reached a conclusion.

The Best Test: Turn AI Off

There’s a simple way to find out whether you’re relying on AI too much.

Stop using it for a while.

You don’t have to abandon AI permanently. Instead, deliberately remove it from a task you’re normally comfortable outsourcing and see what happens.

Write your emails without AI for a day.

Research a topic using traditional search and see whether you can still find and evaluate information yourself.

Write some code without an AI assistant and see whether the logic comes naturally.

Try solving a problem before asking a chatbot for help.

If you discover that a skill has weakened, that’s useful information — not a reason to panic.

Skills can be rebuilt.

Read. Write. Practice arithmetic. Solve problems. Make decisions. Do the things that require you to think.

AI Doesn’t Have to Make You Less Capable

The biggest lesson isn’t that you should stop using AI.

It’s that you shouldn’t outsource every task simply because AI can do it.

Use AI to handle repetitive work. Let it help you explore ideas, provide feedback, explain difficult concepts, and challenge your thinking.

But keep practicing the skills that matter to you.

The difference between healthy AI use and harmful dependence may come down to one simple question:

Is AI helping me think, or am I letting AI think for me?

Use it as a collaborator, not a replacement for your own judgment, memory, creativity, and problem-solving.

AI can save you time.

Used intentionally, it may even help you become better at what you do.