How ChatGPT Slowly Destroys Your Brain: The Alarming Research on AI’s Impact on Learning

How ChatGPT slowly destroys your brain

How ChatGPT slowly destroys your brain is not hyperbole—it’s the alarming conclusion of recent MIT research. If you’ve been using ChatGPT to learn for the last year, you are probably dumber today than you were a year ago. The study “Your Brain on ChatGPT” reveals disturbing findings about how artificial intelligence is fundamentally changing our cognitive abilities, and not for the better.

This isn’t just about productivity or efficiency. It’s about the very foundation of human intelligence and learning. As AI becomes more integrated into our daily lives, understanding its true impact on our brains becomes crucial for anyone who wants to maintain their competitive edge in an increasingly AI-driven world.

The MIT study: Your brain on ChatGPT

Researchers at MIT conducted a groundbreaking study that separated people into three groups and asked them to write essays. The results were both shocking and concerning:

Group 1: LLM Group (ChatGPT users)

This group used only ChatGPT and other large language models like DeepSeek or Gemini. The results were devastating:

• Significantly lower brain activity
• Weaker brain connectivity
• Lower levels of engagement
• Worse information recall
• Poorer quality, more generic work

Group 2: Search Engine Group

This group was allowed to use any website except AI tools. They performed better than the ChatGPT group but still relied on external sources.

Group 3: Brain Only Group

This group used only their own cognitive abilities. They showed the highest levels of brain activity, connectivity, and engagement.

The most concerning finding: Residual damage

Perhaps the most alarming discovery was that even after the LLM group stopped using AI, their brain activity levels didn’t return to normal. This suggests there’s a residual negative effect of using AI that persists even when you’re not actively using it.

This finding is particularly troubling because it indicates that AI use doesn’t just temporarily affect your cognitive abilities—it may be causing lasting damage to your brain’s capacity for deep thinking and learning.

Why AI is making you dumber

To understand why AI is harming our intelligence, we need to examine how learning actually works in the human brain.

The learning process: Information processing

When you receive information from any source, your brain must perform a series of different types of thoughts to organize this information and make it meaningful. This is the crucial step where learning actually happens—called information processing.

When you do this processing well, you develop:

• Better memory
• Deeper expertise
• Stronger critical thinking skills
• Enhanced problem-solving abilities

The cognitive bypassing problem

Traditional learning methods (books, lectures, Google search) require you to actively process information. The information isn’t perfectly packaged for easy consumption, so your brain must work hard to:

• Connect different pieces of information
• Compare and contrast ideas
• Think about the big picture
• Form what researchers call a “schema of knowledge”

With AI tools like ChatGPT, we can skip this crucial processing step. Instead of doing the hard work of organizing and understanding information, we can simply prompt ChatGPT to do it for us. This creates what’s called the “illusion of learning.”

The illusion of learning

The illusion of learning stems from a fundamental misunderstanding: we believe that when we understand something we read, we have learned it. This is absolutely not true.

Just because you can understand something that ChatGPT explains to you doesn’t mean:

• The information will stick in your memory
• You can use that information for complex problem-solving
• You’ve developed real expertise
• You can apply it in new contexts

Real learning requires the effortful processing that AI allows us to bypass. By skipping this process, we’re also bypassing the results—better memory, expertise, and problem-solving ability.

The compounding problem: Skill atrophy

When you get used to bypassing the information processing step, your brain doesn’t develop the habit of processing information effectively. This creates a compounding problem:

If a topic is overwhelming and you use ChatGPT to make it less overwhelming, that topic will always be hard for you because you never develop the skills and thought patterns necessary to organize it yourself.

This means that over time, your ability to learn independently actually decreases, making you more dependent on AI tools rather than less.

The hallucination problem

To make matters worse, AI tools are prone to hallucination—generating text that isn’t necessarily true. If you don’t already have expertise in a domain, you don’t know enough to verify whether the AI’s output is accurate.

In fact, if you’ve used ChatGPT for deep learning or complex problem-solving, it’s almost certain you’ve already learned something that was wrong.

Why AI hallucinates

ChatGPT and other LLMs don’t have access to a universal source of truth. They work through probability networks, determining what word is most likely to come next in a sequence based on their training data.

For complex, contextual problems with limited publicly available information, the AI’s “high probability” answer can be very unreliable. This is why Apple’s recent white paper showed that LLMs are barely capable of reasoning beyond the most basic level.

The expertise paradox

Ironically, AI tools don’t make expertise less important—they make it more important. When you have expertise in a subject, you can:

• Check the AI’s answers for accuracy
• Write better, more specific prompts
• Get higher quality results from your AI interactions

This creates a paradox: AI makes it easier to get generic, mainstream answers, but it also makes it clearer when someone lacks true expertise. The threshold of expertise you need to remain competitive is actually higher than ever.

Real-world example: The data scientist’s struggle

I recently worked with a programmer and data scientist who had spent weeks and dozens of hours in ChatGPT trying to analyze data and figure out the best way to build a dashboard. Despite all this time with AI, when I asked him to explain his approach and strategy at a high level, he really struggled.

He had spent so much time prompting ChatGPT that he never developed his own mental model of the problem. After an hour of discussion, we achieved a level of clarity he had never had before—and he realized he didn’t even need ChatGPT in the first place.

This is a perfect example of how AI can create the illusion of progress while actually hindering real understanding and problem-solving ability.

Why this matters for your career

As AI becomes more capable, the standard for human performance will only increase. Just as knowing how to use Google became a basic requirement rather than a competitive advantage, the ability to think deeply and solve complex problems will become even more valuable.

If AI can provide high value already, why would someone pay you? The answer is that you need to bring expertise and thinking ability that AI cannot replicate. This means:

• Developing deep domain knowledge
• Building strong critical thinking skills
• Maintaining the ability to process information independently
• Creating unique insights and solutions

How to use AI correctly without sacrificing your intelligence

The solution isn’t to avoid AI entirely—it’s to use it as an assistant rather than a replacement for your brain. Here’s how:

1. Use AI for information gathering, not processing

Use AI to:

• Find resources and different perspectives
• Get quick overviews of complex topics
• Act as a sounding board for ideas
• Save time on menial tasks

Avoid using AI to:

• Do your thinking for you
• Organize information you should process yourself
• Generate solutions without understanding the problem
• Replace the effortful work of learning

2. Maintain the processing step

When you feel like learning is taking effort, things are confusing, or information feels overwhelming, that’s a sign your brain is about to start learning. Don’t bypass this feeling—lean into it.

This effortful processing is where real learning happens. Use AI to make the process less tedious, but don’t eliminate the mental work entirely.

3. Challenge and interrogate AI responses

When using AI, focus on:

• Looking for gaps in the information
• Identifying missing perspectives
• Finding connections between ideas
• Challenging assumptions and conclusions

Don’t just accept what AI tells you—use it as a starting point for deeper thinking.

4. Combine AI with traditional learning

Use AI to get an overview, then dive deeper with traditional methods:

• Read journal articles and textbooks
• Practice problem-solving exercises
• Engage in hands-on projects
• Discuss concepts with experts

The future of AI and human intelligence

Some people argue that as AI becomes more capable, the need for human expertise will decrease. This is a dangerous misconception. As AI raises the standard for what’s possible, the threshold for human expertise will only increase.

Just as calculators didn’t eliminate the need for mathematical thinking but made it more important to understand mathematical concepts, AI will make deep thinking and expertise more valuable, not less.

Building your cognitive resilience

To remain competitive in an AI-driven world, you need to:

1. Develop strong information processing skills

Practice the mental work that AI allows you to skip:

• Connecting different pieces of information
• Identifying patterns and relationships
• Building mental models of complex systems
• Developing your own frameworks for understanding

2. Maintain independent learning ability

Don’t become dependent on AI for basic learning tasks. Practice:

• Reading and processing information without AI assistance
• Solving problems using your own reasoning
• Building expertise through traditional methods
• Developing your own unique insights

3. Use AI strategically

Think of AI as a powerful tool that amplifies your existing abilities rather than replacing them. Use it to:

• Accelerate your learning process
• Access information more efficiently
• Explore different perspectives quickly
• Handle routine tasks so you can focus on complex thinking

Live example — user point of view

I used to be someone who relied heavily on ChatGPT for learning. When I encountered a complex topic, my first instinct was to ask ChatGPT to explain it in simple terms. It felt so much easier than struggling through dense textbooks or academic papers.

At first, I thought I was learning more efficiently. I could understand complex concepts quickly, and I felt like I was making rapid progress. But over time, I noticed something troubling: I was becoming less able to think through problems independently.

When I tried to explain concepts to others, I found myself struggling to articulate ideas that I thought I understood. When faced with new problems, I felt lost without being able to ask ChatGPT for help. I realized I had developed a dependency that was actually making me less capable.

The turning point came when I had to work on a project without internet access. I was shocked by how little I could actually do on my own. I had been using ChatGPT so much that I hadn’t developed the mental models and problem-solving skills I needed for independent work.

Now I use AI very differently. I use it to get quick overviews and find resources, but I always do the deep thinking work myself. I read the original sources, work through problems step by step, and build my own understanding. It’s more effortful, but I’m actually learning and retaining information much better.

The key insight was realizing that the feeling of struggle and confusion isn’t a sign that I’m failing—it’s a sign that my brain is doing the work necessary for real learning. By avoiding that work with AI, I was actually preventing myself from developing the skills I needed to succeed.