As artificial intelligence tools become part of daily life, researchers are raising an important question are we relying too much on AI to think for us? A growing number of studies suggest that overuse of tools like ChatGPT, Google Gemini and Claude may affect how people think, remember and solve problems.
What researchers are observing
At the Massachusetts Institute of Technology, researcher Nataliya Kosmyna noticed something unusual. Students were submitting essays that looked very similar. Many also struggled to recall what they had written. This raised concerns that AI tools were being used heavily for thinking tasks. At the same time, researchers observed that people may be forgetting information more easily, possibly due to increased reliance on AI.
Study shows drop in brain activity
To understand this better, researchers at MIT conducted a study with 54 students.
They divided them into three groups:
One used Google Search (without AI summaries)
The results showed a clear difference.
Students who used their own thinking showed strong brain activity across multiple areas. Those using search engines also showed good activity, especially in visual processing. However, the ChatGPT group showed up to 55 per cent less brain activity, especially in areas linked to creativity and deep thinking.
Memory and ownership also affected
The study found that people using AI tools had weaker memory recall.
Could not quote from their own essays
Felt less connected to their work
This suggests that when AI does most of the thinking, people may not fully process or remember the information. Other research has also described this behaviour as “cognitive surrender”, where users accept AI answers without questioning them.
Impact beyond students
The effects are not limited to classrooms.
A separate study found that medical professionals using AI tools for diagnosis became less accurate when they later worked without those tools. Researchers warn that over-reliance on AI could reduce problem-solving ability and independent judgement over time.
Long-term risks for brain health
Experts like neuroscientist Vivienne Ming say the long-term risks could be serious. Low brain activity, especially reduced gamma wave activity, has been linked in some studies to cognitive decline later in life. There are also parallels with earlier technologies. For example, heavy use of GPS has been linked to weaker spatial memory, which is an early indicator in conditions like Alzheimer’s disease.
Can AI still be useful?
Despite these concerns, researchers say AI can still be helpful if used the right way.
A small group in the study (less than 10 per cent) used AI differently. They:
Gathered information using AI
These participants showed stronger brain activity and better results.
Experts suggest a balanced approach:
Think first before using AI
Use AI to challenge ideas, not replace them
Ask AI questions instead of taking direct answers
The way forward: Human + AI
Researchers believe the future lies in “hybrid intelligence”, where humans and AI work together.
Humans focus on thinking and judgement
AI supports with speed and data
The key idea is simple AI should assist thinking, not replace it. As AI tools become more powerful, how we use them will decide whether they improve our abilities or slowly weaken them
