Dear Readers,
What happens in our heads when we work with ChatGPT? A new study from MIT provides a disturbing answer: brain activity decreases when people use AI to write. Fewer neuronal connections, shorter texts, decreasing cognitive effort - a disturbing picture that immediately leads some to ask: Is AI making us dumber?
But perhaps this question is too short-sighted.
After all, every technology makes things easier. The pocket calculator has pushed back mental arithmetic - but has it abolished thinking? Rather, it has made room for more complex tasks. AI could reorganize thinking in a similar way: less mechanical formulation, more strategic thinking - if we let it.
But this is precisely the crux of the matter: we are at a turning point where society must define what “thinking” will mean in the future. Is it the acquisition of knowledge? Combining, questioning, challenging answers - including those of AI?
In Today’s Issue:
Your brain on ChatGPT: MIT scans reveal a surprising cognitive cost.
AI's 'moral compass' might be broken, a new deep-dive reveals.
Ex-OpenAI CTO Mira Murati is back with a new startup aimed at business AI.
The biggest threat from AI isn't job loss—it's that we might stop thinking for ourselves.
And more AI goodness…
All the best,


Does using AI make us dumber?
The TLDR
Google DeepMind has unveiled AlphaEvolve, a groundbreaking AI coding agent that blends Gemini models with evolutionary algorithms to autonomously discover and optimize code. By breaking a decades-old matrix multiplication record, AlphaEvolve marks a turning point—where AI doesn’t just assist with code but pioneers new scientific knowledge.
With a drum roll in your head: What really happens in our brains when we use ChatGPT?
Researchers at MIT investigated how writing with AI affects brain activity. They divided 54 volunteers into three groups: one used ChatGPT, one used Google, and one used no help at all. While everyone wrote essays, their brain activity was measured using EEG. The surprising result: ChatGPT users showed significantly less neural activity, weaker connections between brain regions – and relied more and more on AI the more they used it. The texts became shorter and the cognitive effort less.

This shows that AI can make our thinking easier – but it can also lull us into complacency. Those who use it uncritically leave more and more of the thinking to the machine. The exciting question for the AI community is therefore: How can we manage to see AI as a support without losing our cognitive abilities?

At the same time, however, we must ask ourselves what critical thinking actually is and whether the cause of a lack of critical thinking lies in the education system, even if we take the study mentioned here seriously.
Why it matters: This research shows that AI use measurably changes our brains.
It calls for technology to be designed in such a way that it does not replace but rather strengthens what makes us human. At the same time, however, we must question the study insofar as knowledge and critical thinking are social responsibilities and educational goals. Reducing everything to AI and chatbots is too simplistic!
Sources:
Ad
Find out why 1M+ professionals read Superhuman AI daily.
In 2 years you will be working for AI
Or an AI will be working for you
Here's how you can future-proof yourself:
Join the Superhuman AI newsletter – read by 1M+ people at top companies
Master AI tools, tutorials, and news in just 3 minutes a day
Become 10X more productive using AI
Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.
In The News
AI's 'Moral Compass' is Flawed, Study Finds
In the first large-scale analysis of its kind, researchers have "x-rayed" 10 leading AI reward models (RMs), the systems that act as the moral compass for large language models. The eye-opening results reveal that these crucial safety components are surprisingly flawed, showing wild disagreements with each other, inheriting biases from their base models, and exhibiting other unpredictable quirks.
Mira Murati's New Venture Takes on Business AI
A clearer picture is emerging of the mission for Thinking Machines Lab, the new startup from former OpenAI CTO Mira Murati. According to investors, the highly-funded company is focusing on "RL for businesses," which involves applying reinforcement learning to optimize complex corporate strategies and decision-making processes.
The Threat of Cognitive Offloading
In a recent column, Allysia Finley argues that the biggest threat from AI isn't job displacement but the temptation for young people to engage in "cognitive offloading," weakening their ability to think critically.
Graph of the Day
“The bottlenecks to >10% GDP growth are weaker than expected, and existing $500B investments in Stargate may be tiny relative to optimal AI investment.”

Schrödinger Bridge Matching for Tree-Structured Costs and Entropic Wasserstein Barycentres
This paper introduces a new method for efficiently linking AI models with tree structures and entropy-based transport costs – particularly useful for complex data distributions in bioinformatics or network analysis.
Generative AI for Software Architecture: Applications, Trends, Challenges, and Future Directions
This paper provides a comprehensive analysis of how generative AI such as GPT helps with software architecture – from code generation to decision support. It highlights key trends, tools, and open challenges.
An Empirical Study of GPT‑4o’s Image Generation Capabilities
A systematic investigation of how powerful GPT-4o is at image generation – from text-to-image to 3D models. The study shows how versatile multimodal AI already is today. GPT-4o stands out in particular.
Question of the Day
Do you have the feeling that AI makes you dumber?
Quote of the Day

Sponsored By Vireel.com
Vireel is the easiest way to get thousands or even millions of eyeballs on your product. Generate 100's of ads from proven formulas in minutes. It’s like having an army of influencers in your pocket, starting at just $3 per viral video.