Why leaning on AI can dull the thinking it replaces
The concern is not that AI is harmful in itself but that it can take over mental work the user would otherwise do. When an AI tool produces an analysis, a summary, or a decision, the user can skip the effortful steps — recalling, comparing, questioning — that build and maintain those skills. A review of research on students describes exactly this pattern: overreliance on AI-driven tools may reduce motivation for self-directed analysis, lower cognitive effort, and weaken critical thinking [1]. A separate review frames the same process as cognitive offloading, in which excessive reliance on AI may weaken independent thinking and long-term memory retention [4].
The memory part deserves care. The offloading review reports that some studies find AI supports learning and creativity, while others point to weaker retention when reliance is excessive [4]. That is a mixed picture, not a settled verdict, and it comes from a review of other work rather than a single controlled test. The clearest signal across these studies is about critical thinking under heavy reliance, with memory retention flagged as a related risk rather than demonstrated as a uniform outcome.
What AI literacy seems to change — and what it may not
AI literacy is usually shorthand for knowing how these tools work, where they fail, and when to do the thinking yourself. One study tested whether that knowledge protects students who depend heavily on AI. It found that AI literacy buffered the negative impact of AI dependence on critical thinking — students with stronger literacy kept more of their critical thinking under the same level of reliance [2]. That is the most direct evidence for the idea in the question, and it supports a qualified yes.
The qualification comes from the same study. AI literacy did not simply cancel the cost of dependence; it also amplified cognitive fatigue when AI dependence was high [2]. In plain terms, knowing more about AI may keep you mentally engaged enough to notice problems and second-guess outputs — which is useful for thinking but tiring. So literacy appears to shift the cost rather than erase it: less erosion of critical thinking, more mental strain. Because this finding comes from a single study reported in abstract form, it should be read as a promising mechanism rather than a confirmed trade-off.
The practical answer is balance, not abstinence
Studies of AI-supported classrooms point the same direction. One analysis of attention, reasoning, and higher-order thinking found both benefits and risks, concluding that AI use can harm independent thinking if not used correctly and recommending balance in how much AI students lean on [3]. That recommendation is consistent with the literacy finding: the goal is not to avoid AI but to keep the user doing enough of the cognitive work.
Educators describe the same need from the inside. A qualitative study of medical educators in Palestine, who themselves depend heavily on AI, reported consequences such as reduced critical thinking and decision-making, and argued that AI literacy should be built into the curriculum so students learn how to use these tools [5]. Notably, that study calls for literacy without testing whether it protects memory or independent thinking, so it supports the rationale for teaching AI literacy rather than proving its protective effect.
What we still cannot say for sure
No study here directly measured whether AI literacy preserves memory over time. The memory concern comes from a review noting that excessive reliance may weaken long-term retention [4], and the literacy evidence comes from a study focused on critical thinking and fatigue [2]. Linking the two — literacy protecting memory specifically — is an inference, not a demonstrated result.
The evidence also comes mostly from students in educational settings, often summarized at the abstract level, so it says little about how these effects play out for professionals or over years of use. What the research does support is a modest, conditional claim: heavy AI use is associated with weaker critical thinking, and AI literacy appears to soften that association in at least one study, while possibly adding mental fatigue [1][2]. Whether that adds up to protection for memory and independent judgment in everyday life remains an open question.
Sources used in this answer
Will the Use of AI Undermine Students Independent Thinking?
A review of research on students concludes that overreliance on AI tools may reduce motivation for self-directed analysis, lower cognitive effort, and weaken critical thinking.
Learners' AI dependence and critical thinking: The psychological mechanism of fatigue and the social buffering role of AI literacy
AI literacy buffered the negative impact of AI dependence on critical thinking but also amplified cognitive fatigue when AI dependence was high.
The Cognitive Shift in AI-Supported Education: A Study of Attention, Reasoning, and Higher-Order Thinking Skills Among Students
An analysis of AI-supported education found both benefits and risks to attention, reasoning, and higher-order thinking, concluding that AI can harm independent thinking if not used correctly and recommending balanced use.
Cognitive Offloading in the Age of AI: Are Students Thinking Less or Learning Differently?
A review of cognitive offloading reports that excessive reliance on AI may weaken independent thinking and long-term memory retention, while some studies find AI supports learning and creativity.
Paradox of AI in higher education: Qualitative inquiry into AI dependency among educators in Palestine
A qualitative study of medical educators in Palestine reported that heavy AI dependence was associated with reduced critical thinking and decision-making, and called for integrating AI literacy into the curriculum.