Brain vs Bot: Are We Offloading Too Much To AI?

🧠 Greeting!

• Brain vs Bot: How leaning on AI might be making us mentally lazy—and what that means for students and knowledge workers.

• Poll: Are you sharpening your mind… or letting ChatGPT do the thinking?

• Headlines: Trump’s AI-energy summit sparks $100B+ in U.S. investment; Murati’s startup hits $12B valuation; Nvidia unlocks chip sales to China.

  Let’s dive in. No floaties needed.

The Laboratory

Brain vs bot: Finding focus in the age of AI assistance

Recently, an old report from Reuters caught my eye. The headline read “The Internet will make you smarter, say experts.” The following article went on to say that during an online survey of web users and experts, three-quarters of the people said they believed the internet would make people smarter in the next 10 years. The story was published in 2010. Around the same time, another article from The Wall Street Journal was asking the question, “Does the Internet Make you Dumber?” The article talked about how a growing body of scientific evidence suggested that the internet, with its constant distractions and interruptions, was turning users into scattered and superficial thinkers. These old debates resonate today as we grapple with similar concerns surrounding new technologies.

This time around, the subject of the headlines is artificial intelligence, but the overall question remains the same: “Is AI making us dumb?”

Cognitive risks in the age of ChatGPT

Since the launch of ChatGPT in 2022, companies have rushed to launch more and more powerful generative AI tools to fast-track images, text, and video creation. While this may seem like a great way to improve efficiency, these tools can also be used to offload the cognitive effort that goes into the content creation.

A study by Microsoft and Carnegie Mellon University suggested that AI, while speeding up the completion of menial tasks and boosting productivity, could be making us dumber. Based on a survey of 319 knowledge workers, the study found that increased reliance on AI was associated with decreased self-perceived critical thinking abilities.

The study also found that reliance on AI changed the way workers enacted critical thinking faculties, shifting their focus towards “information verification, response integration, and task stewardship.” When workers had high confidence in AI tools, their perceived effort in completing the task was reduced; however, for workers who were confident in their skills, the perceived effort increased, especially when they were evaluating and applying AI responses. This was not the only study suggesting that relying on generative AI tools may not be in the best interest for cognitive functioning.

MIT study corroborates Microsoft’s findings

Another study, conducted by researchers at MIT’s Media Lab, found something similar. The study was conducted by dividing 54 subjects into three groups and making them write several SAT essays using OpenAI’s ChatGPT, Google’s search engine, and nothing at all. It found that of the three groups, ChatGPT users had the lowest brain engagement and “consistently underperformed at neural, linguistic, and behavioral levels.”

The study further revealed that over the course of several months, ChatGPT users got lazier with each subsequent essay, often resorting to copy-and-paste by the end of the study. In essence, using AI models could harm learning, especially for younger users.

At this juncture, we can safely establish that people who use AI tools for cognitive tasks show lower brain function and engagement, suggesting reduced cognitive function. Additionally, people who used ChatGPT when writing essays felt less ownership of their work and were less likely to recall their work. However, despite the bleak outlook presented by these studies, there is a lot more evidence that needs to be unpacked to understand the larger impact of AI tools on human cognitive functioning.

Reliability and limitations of studies

Studies by Microsoft and MIT’s Media Lab raise concerns about cognitive offloading to AI. However, it’s crucial to note that MIT’s study is yet to be peer-reviewed and involved a small sample size.

Similarly, findings of the study conducted by Microsoft are not foolproof. The study relies on self-reported data, rather than cognitive testing. And historically, humans have not been the best at evaluating themself.

Finding the balance in AI integration

Regardless of whether we like it or not, AI tools are making companies rethink their workloads, resulting in leaner workforces. For others, AI tools are fast becoming part of their workflows to keep up with the pressures of delivering within tight deadlines.

As for the students, AI companies are working hard to entrench their products deep inside classrooms, designing tools that will adapt to students' pace and needs, while assisting teachers to track performance and tailor lessons.

Thus, rejecting AI entirely to safeguard cognitive abilities might not be a practical solution.

That said, just like the body needs exercise and a balanced diet to maintain a healthy lifestyle, the brain needs cognitive challenges or exercise. So, while AI tools may come in handy, excessive reliance on them could lead to brain atrophy, making it exist in co-dependence with artificial intelligence. While this may not sound so bad, the problem arises when the brain is faced with a challenge that cannot be overcome with the help of an AI assistant, like coping with grief, conflict, or uncertainty. So, while AI tools are great as assistants, they cannot be treated as replacements for critical analysis, logical reasoning, and real-world implementation of tasks using self-developed cognitive abilities.

It is essential to remember that although the internet reshaped our attention spans, it didn't diminish the value of storytelling or long-form writing. Instead, it facilitated global engagement among writers and readers, fostering meaningful communities. Similarly, while AI tools march on in their advance towards superintelligence, we will have to draw lines between seeking out their assistance, which is what they are designed to be, and depending on them, which can really turn users into “dumb” machines capable of only copy and pasting things a chatbot spits out.

TL;DR

Quick Poll

Headlines You Actually Need

1. Trump’s AI‑energy summit unleashes $100B+ U.S. investments: Google, Blackstone, CoreWeave, Meta & more commit to power & data‑infrastructure projects around Pennsylvania.

2. Mira Murati’s Thinking Machines valued at $12B: Startup raises $2 b from a16z, Nvidia, Accel & others to build safer, open‑source AI tools.

3. Nvidia clears way for H20 chip sales to China: U.S. links rare-earth trade to export approval, Nvidia stock jumps ~4%.

4. U.S. startup AI funding up 75.6% in H1 2025: PitchBook reports a record $162.8 b went to startups—VC fundraising lagged by comparison.

Rate this edition