A new survey conducted by American technology company GoTo paints a troubling picture of what happens to workers' cognitive abilities when AI tools are used extensively. The findings, reported by Digi.no, suggest that young workers in particular are at risk.
"Cognitive offloading" — when the brain stops working
At the heart of the problem is a phenomenon researchers call cognitive offloading. When workers consistently delegate memory tasks, analysis, and problem-solving to AI systems, the brain can gradually lose its ability to perform those tasks independently. A comparison commonly used in research circles is telling: it is like letting a robot do all your workouts at the gym — the body weakens rather than grows stronger.
It is like letting a robot do all your workouts at the gym — the body weakens rather than grows stronger.
According to the source material underlying the GoTo study, research shows that students and workers who use AI to complete tasks do perform better on the task itself in the moment — but achieve far weaker results on subsequent tests where AI support has been removed. The skills simply do not stick.

Young workers hit hardest
The 17-to-25 age group stands out as especially vulnerable. They use AI tools more frequently than older colleagues, and they simultaneously score lower when critical thinking is measured independently of AI assistance. The researchers behind the study note that higher levels of education appear to act as a partial buffer against these effects — suggesting that the ability to use AI strategically, rather than uncritically, plays a decisive role.
MIT study points to reduced brain activity
The GoTo survey is supported by findings from other research communities. A study from MIT Media Lab — which it is important to note has not yet been peer-reviewed and is based on a limited sample — indicated that participants who were heavily dependent on AI during writing tasks showed lower levels of brain activity and weaker neural connectivity. This was especially pronounced in regions associated with memory, creativity, and executive function.
It is important to stress that research in this field is still evolving. The source material shows that the relationship between AI use and cognition is complex and non-linear. Well-designed AI tutoring systems with active guidance and adaptive support have, for example, been shown to produce better learning outcomes than traditional instruction — suggesting that design and implementation are critical factors.
Not AI, but how it is used
Researchers and psychologists cited in the source material emphasise that the solution is not to reject AI as a work tool, but to use it consciously and strategically. Mindless delegation of cognitive tasks to algorithms is the problem — not the technology itself.
For employers and organisations, the findings raise questions about training and guidelines for AI use, particularly for new hires and young employees who are at a stage where professional skills are still being formed.
GoTo is a technology company that offers AI-integrated business tools such as GoTo Connect and GoTo Contact Center, and therefore has a clear commercial interest in the subject. There is good reason to read the company's own surveys with a critical eye — even if the findings align with independent research in the field.
