In one of my recent presentations on “AI in the digital society”, I already talked about what I call “the GenZ dilemma”. The idea goes like this: GenZ is the first generation growing up in the new zettabyte era. This generation has a very high “digital tool adoption”, but that isn't enough to reach the necessary “digital maturity”, meaning the skills you actually need in this digital age.
A new MIT study is currently making the rounds. It highlights the problem in relation to the new AI tools and confirms a trend that earlier studies had already pointed to. These tools make us “mentally lazy” and get in the way of building important cognitive abilities. And this hits hardest those who are right in the middle of their learning process*, in other words GenZ.
First, a summary of the study:
The participants were split into three groups:
- AI group: allowed to use ChatGPT (GPT-4o) only.
- Search engine group: could use any websites, but no AI.
- Brain group: worked without any tools at all.
Across several sessions, all participants wrote essays on various topics while their brain activity was recorded via EEG. On top of that, the texts were graded by AI and by human teachers, the participants were interviewed, and their ability to quote content from their own essays was tested.
The results are clear:
- Brain activity: The brain group showed the strongest and most widespread neural connectivity. The search engine group landed in the middle, while the AI group showed the weakest coupling between brain regions.
- Memory and recall: Participants who worked with AI had serious trouble quoting content from their own essays. They remembered less and felt less connected to what they had written.
- Ownership: Those who wrote exclusively with AI reported identifying less with their own text. The brain group and the search engine group felt much more like actual authors.
In a follow-up session, some participants switched methods:
- From AI to brain: Those who had previously worked with AI kept showing weak neural connectivity and less engagement after switching to “brain only”. Worse still, this group also tends to copy the writing style of the AI they had used; the study calls it a “contamination effect”
- From brain to AI: Those who first wrote without any tools and then switched to AI were better able to maintain their brain activity and memory performance, and showed broader neural activation.
In short
- AI makes writing more efficient, but also more superficial.
- Creativity, memory and ownership suffer when AI is used.
- Anyone who works with AI regularly risks their own thinking becoming shallower and less connected.
- Targeted, deliberate use of AI – for example after a phase of independent thinking – can soften the downsides.
AI is a powerful tool that can speed up and simplify learning. But the price of this efficiency is high: lean on AI too much, and you lose creativity, memory and any sense of ownership of your own work.
A bit of neurobiology to finish
None of this is really surprising: our brain works like a muscle. Train it and it gets stronger. Underuse it and it wastes away. While older generations spent years training and strengthening their brains through school, university and work, GenZ is right in the middle of that development. And that development can be seriously disrupted by an unbalanced use of AI. When you consider that this disruption hits precisely the "skills" employers will expect in 2030 (see the end of Episode 10.3), e.g. "creative thinking" or "analytical thinking", you may understand why I so often talk about the need for "social change management" when it comes to AI adoption. This problem has to be tackled in education and training now. Not later.
Bottom line:
- AI as an assistant for GenX through Boomers? Brilliant!
- AI as an unsupervised learning buddy for GenZ (and younger)? Disaster!
A few sources
- The Impact of Generative AI on Critical Thinking
- Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task
Live long and prosper 😉🖖
