Artificial intelligence has woven its way into everyday life, prompting questions about how reliance on generative tools affects the human mind. A recent small study published in Applied Cognitive Psychology suggests that leaning heavily on ChatGPT can reduce critical thinking, diminish memory, and lead to social isolation. However, users who engage with the software intentionally can also harness it to boost their creativity and reasoning skills.
Over the last few years, tools that generate human-like text have become wildly popular. ChatGPT reached one million users within a week of its launch and currently boasts over 180 million active users. People use the program for everything from writing emails to solving programming problems.
As these programs become a regular part of daily workflows, researchers are trying to understand the psychological and cognitive costs. Zeineb Farhat, a researcher at ISC Paris, initiated the research to investigate what happens when habitual use turns into dependence. Much of the early research on generative software has focused on its benefits in education and the workplace.
Farhat wanted to document the potential downsides of outsourcing mental effort and emotional needs to a machine. The research draws on established ideas in psychology regarding how people manage mental effort and motivation. One concept is cognitive offloading, which happens when individuals use external tools, like calculators or search engines, to reduce the amount of information they need to process or remember.
Another core idea involves the psychological needs for autonomy, competence, and social connection. When people feel capable and in control of their actions, they are generally more motivated. Farhat set out to see how constant access to an intelligent digital assistant might alter these processes.
The researcher wanted to know if offloading cognitive tasks to an algorithm would compromise a user’s ability to think independently. To explore these dynamics, Farhat conducted a small study involving 45 young adults in France. The participants, aged 18 to 25, were all regular users who interacted with ChatGPT daily.
The researcher conducted semi-structured interviews with each participant, either face-to-face or via video call. These interviews lasted between 45 and 60 minutes. The conversations covered how the participants used the software, how it affected their study or work habits, and how it made them feel.
Farhat then analyzed the interview transcripts to identify recurring themes in the users’ experiences. The analysis revealed that heavy reliance on the software often discouraged independent thought. Participants reported that having instant answers at their fingertips made them less likely to analyze problems deeply or evaluate sources.
Some described experiencing intellectual laziness, noting that their attention spans felt shorter. Users also reported trouble retaining information. They were externalizing their memory by relying on the chatbot to retrieve facts rather than practicing active recall.
“I used to take notes and actively review materials, but now I just ask ChatGPT whenever I need something,” one student noted in the study. “The problem is, I don’t retain much. I feel like my memory has become weaker because I rely on ChatGPT instead of trying to recall things on my own.”
Psychologically, the interviews indicated that heavy use eroded the participants’ sense of autonomy and self-trust. Many users started second-guessing their own judgment. They began seeking the software’s validation for simple decisions, like choosing the right words for a basic email.
The ease of getting quick answers also lowered their intrinsic motivation to learn new things. Without the struggle normally required to master a topic, the learning process felt less rewarding. Some participants even reported anxiety and a sense of paralysis when the software was offline for maintenance.
One participant recalled a time when the system went down, bringing their workflow to a complete halt. “I needed it for a work project, and I felt completely paralyzed without it,” the user shared. “It’s like my safety net is gone.”
The study also uncovered shifts in social behavior. Participants frequently chose to interact with the software instead of discussing ideas with peers or teachers. This preference sometimes led to feelings of social isolation, as users lost the habit of engaging in deep conversations with other people.
A few participants began to treat the text generator as a companion, projecting human traits onto it. “ChatGPT has become my friend; it knows everything about me,” one participant stated. “It’s always there, always ready to listen, and it feels like it understands me better than most people do.”
While this offered temporary comfort, it substituted genuine human connection with a machine response. Relying on an algorithm for emotional support risks blurring the line between technological mediation and true social interaction. Over time, this dynamic can weaken interpersonal skills and empathy.
Despite these negative reports, the study identified a pattern of constructive use. Some users treated the software as a collaborative tool rather than a replacement for their own brains. By allowing the program to handle tedious tasks like structuring data or summarizing dense text, these individuals freed up mental energy for complex reasoning.
They used the tool to brainstorm or overcome writer’s block. Instead of copying what the program generated, they used the output as a starting point to think differently. When used reflectively, the software actually increased users’ confidence and reduced their anxiety about challenging tasks.
Because the research relies on qualitative interviews with a small group of young adults in France, the findings reflect a specific demographic and cannot be broadly generalized. People from older generations or different cultural backgrounds might interact with the software differently. The study also captures user experiences at a single point in time.
It remains unknown whether these cognitive and social habits change over years of use. Future research will need to track larger and more diverse groups over longer periods to see if these patterns persist. Additional studies could also explore how to best train people to use these tools adaptively rather than dependently.
The study, “When AI Thinks for Us: Unveiling the Cognitive and Social Toll of ChatGPT Over-Reliance,” was authored by Zeineb Farhat.