The Hidden Costs of Relying on ChatGPT for Learning

AI tools like ChatGPT promise a world of instant knowledge and effortless productivity. With the ability to generate summaries, write essays, and answer complex questions in seconds, it’s no surprise that many turn to these technologies for help in learning. However, recent research from MIT highlights a disturbing trend: overuse of AI tools could be making us less intelligent, not more.

In a study detailed in Your Brain on ChatGPT, researchers examined how people learn and process information when using large language models (LLMs) like ChatGPT, DeepSeek, or Google’s Gemini. The findings reveal significant cognitive downsides when users become dependent on such tools for learning and problem-solving. Here’s what the study uncovered and why it’s essential to rethink how we engage with AI in education and work.

---

Comparing Learning Methods: AI, Search Engines, and the Human Brain

To understand the impact of AI on learning and cognition, researchers divided participants into three groups and asked them to write essays under different conditions.

• LLM Group: This group used ChatGPT and similar AI tools to assist with their writing.
• Search Engine Group: Participants were allowed to use any non-AI tools, such as search engines, to gather information.
• Brain-Only Group: These individuals relied solely on their knowledge and handwriting to create essays.

By employing EEG analysis to monitor brain activity, the researchers discovered differences in how each group processed information:

Group	EEG Findings	Essay Quality	Memory Recall
LLM Group	Lower brain activity and connectivity; weaker engagement.	Poor; generic content	Weakest
Search Engine Group	Moderate brain activity and engagement.	Better	Stronger
Brain-Only Group	Highest brain activity and robust cognitive engagement.	Best overall	Strongest

The study revealed that the reliance on AI bypasses the crucial cognitive effort required to adequately process and internalize information, leaving users with poorer memory and lower-quality work than those who engaged their brains directly. Disturbingly, even after participants in the LLM group stopped using AI tools, their brain activity and cognitive performance showed residual impairments compared to the other groups.

---

Cognitive Offloading: How AI Bypasses Critical Thinking

Learning is an active process that occurs when the brain organizes, evaluates, and connects new information to existing mental frameworks — a process called information processing. This critical step requires effort and mental exertion, but it lays the foundation for memory retention and expertise.

With tools like ChatGPT, however, this cognitive effort is often bypassed. Instead of engaging in the complex work of constructing mental connections and schemas, users let the AI do it for them. The result is an "illusion of learning," where the user mistakes surface-level understanding for true knowledge.

Just because you understand an AI-generated response at a glance doesn’t mean you’ve internalized its substance. Without the practice of critical thinking and active engagement, your memory weakens, and your ability to tackle complex, contextual problems diminishes over time.

Residual Effects of AI on the Brain

The study also found that dependency on AI tools like ChatGPT leads to long-term cognitive repercussions. Once users get into the habit of skipping the mental effort needed to process information, their capacity for deep thinking erodes. Even when they stop relying on AI, they may find it much harder to develop the mental muscle needed to approach complex topics.

---

The Challenge of AI Hallucination

Another critical concern is "AI hallucination," where language models like ChatGPT generate information that seems plausible but is factually incorrect. This issue arises because these models don’t know the "truth"; they merely predict the most probable sequence of words based on their training data.

For users unfamiliar with a given subject, this poses a significant risk. Without a strong foundation of knowledge, they may not have the ability to verify whether the answers generated by AI are accurate. In such cases, reliance on AI tools compounds the problem, leaving learners to internalize incorrect or incomplete information.

---

How to Use AI Tools Without Losing Your Edge

Despite these drawbacks, the video emphasizes that AI tools like ChatGPT aren’t inherently harmful. When used wisely, they can serve as powerful assistants that save time and provide perspectives, making the learning process more efficient. The key is to avoid using AI to replace your own critical thinking. Here are some strategies for leveraging AI effectively without sacrificing cognitive growth:

1. Use AI for preliminary research: Let AI help you get a general overview of a topic, compile basic facts, or find additional resources.
2. Ask better questions: Treat AI as a tool to test your existing knowledge or gain new perspectives, not to replace your thinking.
3. Don’t bypass careful processing: Resist the temptation to accept AI’s answers at face value. Take the time to process, analyze, and verify the information.
4. Focus effort on critical thinking: Engage your brain to connect ideas, identify gaps, evaluate competing interpretations, and build a solid foundation of understanding.
5. Avoid dependency: If you find a particular topic or task difficult to process, consider that discomfort an opportunity to strengthen your mental capabilities. Skills improve through deliberate effort, not shortcuts.

---

Adapting to a World Dominated By AI

As AI technology improves, some people hope for a future where these systems become so accurate and intuitive that individual expertise becomes less necessary. However, the presenter argues that the opposite may hold true.

The more AI develops, the more it raises our standards for human expertise. Using Google effectively was once seen as a skill; today, it’s a basic requirement. Similarly, as AI becomes adept at providing high-quality, mainstream answers, the true differentiators in the job market will be individuals capable of applying deep, critical thinking and tailoring AI tools to solve nuanced problems.

Employers will favor those who can work both with and beyond the capabilities of AI, bringing unique expertise. To remain relevant, individuals must focus on developing critical thinking skills, retaining their knowledge, and using AI as a complement to their learning rather than a replacement for effort.

---

Practical Lessons for Using AI Wisely

The key takeaway is this: AI tools like ChatGPT are not inherently harmful, but their misuse or overuse can lead to significant losses in cognitive abilities and critical thinking. To avoid counterproductive outcomes, consider these practical tips:

• Reframe AI’s role: Think of AI as a tool or assistant, not a replacement for your brain.
• Learn the basics first: Build foundational knowledge to effectively recognize and verify AI-generated information.
• Engage with the process: Don’t skip over the hard, effortful work of learning. Struggles and confusion are necessary for building expertise.
• Diversify your sources: Use AI to pull resources and perspectives but make sure to cross-reference with credible sources.
• Keep practicing deep work: Manual tasks like reading textbooks, taking notes, and reorganizing information improve your ability to think deeply.

---

AI might seem like the ultimate shortcut to learning, but it comes with risks that can hinder your personal and professional growth. By using AI judiciously and dedicating time to internalizing information deeply, you can unlock its potential without falling for the illusion of learning. At the end of the day, cognitive effort is the cornerstone of true expertise.

---