Daily interaction with neural-network assistants may be quietly reshaping how the human brain works. French researchers, who published their work in the journal Applied Cognitive Psychology, concluded that excessive reliance on AI assistants is associated with impaired memory and increased social isolation. This is reported by RBC-Ukraine, citing the study's findings.
What exactly the researchers studied
The study was authored by researcher Zeinab Farhat from the ISC Paris business school. The survey included 45 young people aged 18 to 25 who use ChatGPT daily — for learning, work, or personal needs. The sample was deliberately limited to a single age group, which, as the authors themselves note, limits the ability to generalize the results to a broader audience of users.
"Cognitive offloading": how AI displaces one's own thinking
The key term recorded in the study is so-called "cognitive offloading" — a state in which a person fully delegates the storage and processing of information to an external tool. Instead of retaining knowledge in their own memory and spending effort analyzing it, the user relies on the chatbot. According to the study, it is precisely this habit of "switching off" one's own cognitive effort that is associated with impaired memory among active users.
Social isolation as a side effect
Beyond the cognitive consequences, the scientists also recorded a social effect. Some active users preferred interacting with the chatbot over live discussions with friends, teachers, or colleagues. Gradually, this substitution of real contact with synthetic contact led to feelings of loneliness and social isolation, which, according to the authors, may also exacerbate general self-doubt.
Not all is gloom: selective use
At the same time, the results are not unambiguously negative. Participants who used ChatGPT selectively — as a tool for brainstorming and finding inspiration, rather than as a replacement for their own analysis — conversely reported increased self-confidence and improved creativity. Thus, it is the nature of use, rather than the mere fact of turning to AI, that serves as the key factor influencing cognitive and social well-being.
Contradictory data
In the public coverage of the work, positions diverge. On the one hand, a number of outlets present the study as proof that ChatGPT "kills thinking skills and dulls creativity," emphasizing the risks of cognitive offloading. On the other hand, the authors themselves stress that creativity and self-confidence increased among those who used the assistant in moderation, rather than as a substitute for their own analysis. Moreover, due to the small sample (45 people) and the restriction to a single age group, the conclusions cannot be extrapolated to all users. The difference between the phrasings "AI harms thinking" and "AI harms when over-relied upon" is substantial, and it is precisely this that determines how accurate the sensational headlines are.
Limitations and a call for new research
The authors honestly point out the methodological limitations: the small sample size and the narrow age range do not allow the results to be generalized to the entire population of AI users. The scientists call on the scientific community to conduct larger-scale and longer-term studies of the impact of neural-network assistants on the human psyche, in order to distinguish durable effects from short-term and situational ones.
