Illusion of Competence and Skill Degradation in Artificial Intelligence Dependency among Users

Illusion of Competence and Skill Degradation in Artificial Intelligence Dependency among Users

Dr. Ramzy Muorwel Matueny*, Dr. Joseph Juma Nyamai

College of Health Sciences, Mount Kenya University, Thika Main Campus, Kenya

*Corresponding author

DOI: https://doi.org/10.51244/IJRSI.2025.120500163

Received: 16 May 2024; Accepted: 20 May 2025; Published: 18 June 2025

ABSTRACT

This conceptual review paper explores the emerging phenomenon of skill degradation in the context of increasing reliance on artificial intelligence (AI) within higher education and professional environments. As AI tools become integral to learning, writing, assessment, and decision-making processes, many users particularly students and instructors experience what has been termed the illusion of competence, a misleading perception of mastery created by AI-generated outputs that mask underlying cognitive deficits. Drawing on Cognitive Load Theory and Technological Dependency Theory, this paper examines how the offloading of intellectual effort to AI systems diminishes core human faculties such as memory retention, critical thinking, metacognitive awareness, creativity, and professional judgment. The analysis is structured across multiple dimensions, including the cognitive mechanisms of skill loss, real-world manifestations in academic and occupational settings, and the broader psychological and social consequences of overdependence. It highlights risks such as academic underperformance, reduced originality, erosion of self-efficacy, widening equity gaps, and the devaluation of human expertise. Ethical and pedagogical concerns such as fairness, transparency, data privacy, and faculty readiness are also addressed. The paper concludes with strategic recommendations for educational institutions, including the need for AI literacy training, faculty development, assessment reform, and policy frameworks that encourage responsible and critical engagement with AI technologies. Ultimately, the paper argues for a balanced, human-centered approach to AI integration, one that positions AI as a support system rather than a substitute for cognitive engagement, ensuring that technological advancement enhances rather than displaces the human capacity for deep, reflective learning.

Keywords: Artificial Intelligence, Skill Degradation, Illusion of Competence, Artificial Intelligence dependency

INTRODUCTION

In contemporary society, Artificial Intelligence (AI) has rapidly transitioned from a niche technological advancement to an integral component of daily personal, educational, and professional activities (Russell & Norvig, 2021). AI refers to computational systems capable of performing tasks typically associated with human intelligence, including problem-solving, decision-making, language processing, and data analysis (Nilsson, 2019). The proliferation of AI-driven tools ranging from virtual assistants like Siri and Alexa to sophisticated academic platforms such as Grammarly, ChatGPT, and AI-powered research summarizers has profoundly reshaped how individuals approach tasks previously dependent solely on human cognition (Luckin & Cukurova, 2019).

This transformative integration of AI tools into everyday tasks has given rise to what is increasingly recognized as the “illusion of competence.” This concept describes a deceptive state wherein individuals perceive themselves as capable or skilled due to the frequent reliance on AI tools, whereas, in reality, their actual cognitive and practical abilities might be diminishing or underdeveloped (Bjork & Bjork, 2020). The convenience and immediacy provided by AI assistance may inadvertently discourage active mental engagement, promoting a passive, superficial interaction with tasks that traditionally required rigorous intellectual involvement (Carr, 2020).

The primary aim of this paper is to critically analyze how excessive reliance on AI technology can lead to significant skill degradation, specifically targeting fundamental cognitive skills such as memory retention, analytical problem-solving, and critical thinking capabilities (Greenfield, 2015; Sparrow, Liu, & Wegner, 2011). While AI unquestionably offers substantial benefits, including efficiency and ease of access to information, its unchecked use can undermine the very cognitive abilities it seeks to complement (Kirschner & De Bruyckere, 2017). This nuanced perspective is essential for educators, policymakers, technologists, and students alike, as it prompts a critical evaluation of the long-term implications of integrating AI into daily cognitive tasks (Selwyn, Hillman, Eynon, Ferreira, & Knox, 2020).

The relevance and urgency of addressing AI-induced skill degradation cannot be overstated. Educational institutions and workplaces increasingly integrate AI, often with minimal consideration for its psychological and cognitive consequences (Cukurova, Bennett, & Abrahams, 2018). Understanding and mitigating the negative impacts of AI dependency are essential for maintaining robust intellectual and professional competencies. Without careful management, society risks creating generations of learners and professionals who may find themselves ill-equipped to function effectively in contexts where AI support is unavailable, insufficient, or unreliable (Selwyn et al., 2020; Luckin & Holmes, 2016).

Consequently, this paper will explore these concerns systematically, starting with conceptual clarification and subsequently examining practical implications, consequences, and strategic solutions to manage and balance AI’s role effectively in promoting genuine competence rather than perpetuating an illusion thereof.

Conceptual Background

To comprehensively address the issue of skill degradation due to AI dependency, it is critical first to clarify key concepts and theoretical foundations underpinning this phenomenon.

Key Concepts

Artificial Intelligence

Artificial Intelligence (AI) is broadly defined as computer systems or algorithms capable of performing tasks that typically require human intelligence, such as learning, reasoning, problem-solving, perception, and natural language processing (Russell & Norvig, 2021). The scope and sophistication of AI applications range from simple automation tools to highly advanced generative models capable of simulating human creativity and insight.

The illusion of competence refers to a psychological phenomenon where individuals mistakenly believe they possess certain skills or knowledge because they rely heavily on external aids or technologies (Bjork & Bjork, 2020). This concept closely relates to cognitive biases like the Dunning-Kruger effect, where limited self-awareness leads individuals to overestimate their own abilities (Kruger & Dunning, 1999). The increased availability and integration of AI tools can exacerbate this phenomenon by providing easy access to solutions, thereby falsely inflating a user’s self-perception of competence without actual skill acquisition.

Skill degradation

Skill degradation involves the deterioration or loss of proficiency in previously mastered skills, which occurs through disuse, lack of practice, or overreliance on technological support (Arthur et al., 1998). In the context of AI, skill degradation specifically pertains to the erosion of fundamental cognitive abilities including memory retention, analytical thinking, and problem-solving – those individuals cease to engage actively when they consistently delegate cognitive tasks to AI systems (Greenfield, 2015).

Theoretical Frameworks

This paper utilizes two primary theoretical lenses to understand the relationship between AI dependency and skill degradation:

The Cognitive Load Theory (CLT)

Cognitive Load Theory, proposed by Sweller (1988), postulates that human working memory has limited capacity for processing new information. According to CLT, optimal learning occurs when cognitive resources are efficiently allocated, minimizing extraneous load while maximizing relevant cognitive activities (Sweller, Ayres, & Kalyuga, 2011). AI-driven tools initially seem beneficial by reducing cognitive load; however, prolonged reliance on such tools can result in diminished cognitive engagement, effectively minimizing the mental effort necessary for meaningful learning and skill retention. Over time, learners may lose essential cognitive skills due to insufficient active processing of information (Kirschner & De Bruyckere, 2017).

Figure 2.1: Cognitive Load Theory and the Impact of AI Tools on Learning

Sources: Author’s construction from the literature

The Dependency Theory (Technological Dependency)

Dependency Theory in technological contexts suggests that sustained reliance on specific technologies can create psychological and practical dependency, wherein users gradually lose the capacity to function effectively without these technological supports (Parasuraman & Riley, 1997). In the educational domain, excessive reliance on AI-based tools may foster dependency, thus weakening students’ and professionals’ intrinsic capabilities to perform cognitive tasks independently. This theory provides a lens through which the risks of over-dependence on AI and subsequent skill degradation can be systematically analyzed and understood (Carr, 2020).

In synthesizing these theoretical perspectives, this paper underscores that while AI’s ability to streamline complex tasks can be beneficial, it also poses significant risks. The unintentional consequence of persistent AI integration might be the weakening of fundamental cognitive competencies necessary for sustained academic, professional, and personal development (Sparrow, Liu, & Wegner, 2011).By clearly delineating these conceptual foundations, the subsequent sections of this paper will critically assess practical examples of skill degradation arising from AI dependency, the broader implications of these phenomena, and strategies to mitigate these risks effectively.

Figure 2.2: Technology Dependency and AI in Education

Sources: Author’s construction from the literature

MECHANISMS OF AI-INDUCED SKILL DEGRADATION

As AI becomes more deeply embedded in learning environments, professional settings, and everyday life, it is crucial to examine the specific mechanisms through which this technology can lead to the gradual erosion of essential human cognitive and practical skills. This section outlines how AI use affects cognitive engagement, identifies the particular skills at risk, and highlights the subtle progression from assistance to dependency.

Cognitive Disengagement and Mental Automation

One of the core mechanisms of AI-induced skill degradation is cognitive disengagement, where individuals outsource mental effort to AI systems, thereby reducing the depth of their own cognitive processing. According to Cognitive Load Theory, when tasks are simplified excessively through automation, learners are deprived of the necessary mental challenges that stimulate long-term learning and memory consolidation (Sweller et al., 2011). AI tools often handle complex functions such as summarizing texts, generating ideas, or solving equations, leaving users in a passive role of consuming rather than processing or producing knowledge.

Mental automation occurs when users become habituated to AI-generated answers, diminishing their inclination to verify, analyze, or critically engage with the content. This mental shortcut may increase efficiency in the short term but contributes to a long-term decline in independent problem-solving abilities (Kirschner & De Bruyckere, 2017).

Impact on Memory Retention

The externalization of memory storing knowledge in machines rather than the brain is another critical mechanism of skill degradation. Sparrow, Liu, and Wegner (2011) termed this phenomenon the “Google Effect,” which refers to the tendency of people to forget information that they believe will be readily available through external sources. When individuals habitually turn to AI for facts, concepts, or instructions, they train their minds to retrieve from the machine rather than internalize the knowledge themselves.

This pattern undermines long-term memory development, which is essential for deep learning, creativity, and the ability to apply knowledge in novel situations. Students who consistently rely on AI for answers may struggle with recall during assessments or practical applications, where AI tools may not be permissible.

Erosion of Critical Thinking and Problem-Solving Skills

AI tools like ChatGPT and other generative models are designed to provide fluent, coherent, and seemingly accurate responses. While this can assist users in understanding complex topics, it also presents the risk of bypassing critical thinking. Users may accept AI-generated content at face value without scrutinizing its accuracy, underlying assumptions, or logic.

In academic settings, this undermines Socratic questioning, evidence evaluation, and argument construction – skills that are foundational to intellectual development. When students default to AI instead of constructing their own arguments or analyzing diverse viewpoints, they lose opportunities to strengthen cognitive flexibility and intellectual resilience (Selwyn et al., 2020).

Diminished Writing and Communication Competencies

Writing is a cognitively demanding task that requires organization, grammar control, logical sequencing, and clarity. With the increasing availability of AI writing assistants such as Grammarly, QuillBot, and ChatGPT, users may experience improvements in surface-level writing quality but a simultaneous decline in the underlying skill of composing ideas clearly and independently.

Studies suggest that when learners use AI to reformulate or generate entire sections of writing, they may become less capable of expressing themselves effectively without such tools (Luckin & Holmes, 2016). The overuse of AI for academic essays, reports, and emails can result in a loss of authorial voice, decreased linguistic creativity, and an over-reliance on templated expressions.

Loss of Metacognitive Skills

Metacognition, the awareness and regulation of one’s own thought processes, is a key factor in effective learning. AI tools that provide immediate solutions or automated feedback can interfere with self-monitoring and reflection, core components of metacognitive development. If students depend on AI to highlight errors or suggest improvements without understanding why, they may become passive consumers of feedback rather than active participants in the learning process (Bjork & Bjork, 2020).

This can create a vicious cycle whereby students become more dependent on AI tools to validate their understanding, and their intrinsic motivation to assess or correct their own work diminishes. Over time, this undermines autonomy and self-regulated learning, which are essential for academic and professional success.

Professional Implications: Deskilling in the Workplace

AI overdependence is not confined to education. In professional environments, excessive automation can lead to deskilling, where human workers lose competencies due to the consistent delegation of tasks to AI systems. For example, automated diagnostic tools in healthcare, financial forecasting software in business, or legal research AI in law firms can lead professionals to underutilize their own expertise (Parasuraman & Riley, 1997).

While these tools enhance efficiency, they also risk atrophying judgment, intuition, and domain-specific reasoning. This can be dangerous in high-stakes situations where machine outputs may be flawed, biased, or misinterpreted. If professionals lack the skill or confidence to challenge AI-generated outputs, organizational errors and ethical oversights become more likely.

Summary table 8.3: Key Mechanisms

Mechanism	Description	Consequences
Cognitive disengagement	AI reduces need for mental effort	Surface-level learning, passive habits
External memory dependence	Reliance on AI to store or recall information	Weakened long-term memory
Reduced critical thinking	Acceptance of AI-generated answers without scrutiny	Loss of reasoning and analytical depth
Writing skill erosion	Overuse of AI for content generation and editing	Loss of expressive and linguistic skills
Metacognitive skill decline	AI interrupts self-monitoring and self-correction processes	Lower academic autonomy
Workplace deskilling	Delegation of expert tasks to AI in professions	Diminished judgment and responsibility

Source: Author’s construction from the literature

EXAMPLES OF AI DEPENDENCY IN DIFFERENT CONTEXTS

AI is no longer a futuristic concept, it is a present reality, embedded in academic institutions, professional workplaces, and the daily routines of ordinary users. While its advantages are significant, its subtle shift from a support system to a dependency tool is observable across various domains. This section explores concrete examples of how AI dependency manifests in different contexts, leading to the gradual weakening of essential human capabilities.

Educational Settings: The Student Experience

AI in Writing and Assignments

With the rise of AI writing tools such as ChatGPT, Grammarly, Jasper AI, and QuillBot, students increasingly use these platforms for composing essays, reports, and even answering discussion prompts. On one hand, these tools enhance grammar, coherence, and fluency. On the other hand, the automation of thought reduces opportunities for students to struggle with language and content which is an essential process in the development of academic and intellectual skills (Kirschner & De Bruyckere, 2017).

Instead of learning how to construct a persuasive argument or analyze a text deeply, students may opt to prompt an AI to “write an argumentative essay on climate change,” and merely edit the result. This fosters a sense of completion and competence, but the actual cognitive engagement is superficial, often limited to reviewing AI-generated ideas rather than generating and defending original ones.

AI for Research and Summarization

Students now use tools like Scholarly, Elicit, and Semantic Scholar AI assistants to scan through large volumes of academic literature. These tools highlight key findings, extract research questions, and summarize methodologies. Although time-saving, they can create an illusion of familiarity with complex material. Students may assume they understand an article simply because they read the AI summary, without engaging with the nuanced arguments, limitations, or theoretical implications discussed in the full text (Sparrow et al., 2011).

Exam Preparation and Study Support

AI tutors such as Quizlet AI, Khanmigo, and ChatGPT-based Q&A bots allow students to practice and revise quickly. But when used excessively, they risk promoting rote learning over reflective learning. Students may memorize AI-generated responses without grasping underlying principles or practicing problem-solving under pressure. This results in inflated confidence during preparation, followed by underperformance in assessment scenarios requiring independent thinking.

Educators and Instructional Automation

Lesson Planning and Curriculum Design

Educators are increasingly using AI tools to develop lesson plans, quizzes, and even lecture scripts. Generative AI like ChatGPT or tools such as ScribeSense and Kahoot AI can create instructional materials within minutes. However, this convenience risks undermining pedagogical intentionality, the deliberate alignment of learning objectives, activities, and assessments.

When educators depend on AI-generated lessons without critically adapting content to learners’ needs, contexts, and capabilities, it can lead to standardized and decontextualized teaching, especially in diverse classrooms (Luckin & Holmes, 2016). Over time, this can deskill educators, reduce innovation in teaching strategies, and create disconnects between instruction and real student learning challenges.

Automated Grading and Feedback Systems

AI-based grading tools like Grade scope, Turnitin Feedback Studio, and Writable support fast grading and consistent rubric-based evaluation. However, they also limit holistic evaluation—the teacher’s nuanced judgment of student intent, effort, and learning trajectory.

Overdependence on these systems may desensitize instructors from identifying subtle learning difficulties, creative approaches, or ethical dilemmas in student work. Feedback may become templated and impersonal, which in turn reduces student engagement with and trust in instructor guidance (Selwyn et al., 2020).

Workplace Dependency and Professional Deskilling

Health-care: Diagnostics and Clinical Judgment

AI systems such as IBM Watson for Oncology, Aidoc, and Tempus are now used to support diagnostic imaging, treatment suggestions, and patient monitoring. These tools can help physicians detect patterns or anomalies they might miss—but they also pose risks of clinical deskilling.

When doctors depend on AI recommendations without cross-verifying or critically assessing them, it can erode diagnostic acumen, especially among younger or less experienced practitioners (Cabitza et al., 2018). Medical education risks turning into a protocol-following exercise, with diminished emphasis on critical case-based reasoning or ethical deliberation in complex patient scenarios.

Legal Practice and Research

AI in law, such as ROSS Intelligence or Casetext CoCounsel, is used for legal document analysis, case law retrieval, and drafting motions. These tools significantly cut down research time, but they also threaten deep legal reasoning, which often requires contextual understanding, precedent comparison, and interpretive creativity.

Junior lawyers depending on AI may grow accustomed to accepting AI’s framing of legal issues rather than developing their own interpretations or spotting novel legal arguments—skills crucial to advocacy and jurisprudence (Susskind, 2019).

Journalism and Media

AI platforms like Jasper, Wordsmith, and ChatGPT can generate articles, captions, and news summaries. Some media outlets already use AI for automated news reporting on sports and finance. However, reliance on these systems undermines investigative journalism, originality, and fact-checking. Journalists may gradually lose the instinct for inquiry and source verification if they rely too heavily on AI-generated drafts (McGregor, 2020).

Everyday Life: Cognitive Offloading and Habitual Dependence

Navigation and Wayfinding

GPS apps like Google Maps and Waze have revolutionized transportation, but also contributed to the decline of spatial cognition. Studies have shown that frequent GPS users demonstrate lower levels of spatial awareness and route memory (Ruginski et al., 2015). Individuals increasingly rely on visual or voice prompts for directions—even to familiar places—demonstrating a form of digital dependency that replaces traditional map-reading and environmental scanning.

Daily Tasks and Personal Planning

Digital assistants like Siri, Alexa, and Google Assistant handle reminders, schedules, and to-do lists. While convenient, they also replace self-regulation and memory cues. Many users struggle to recall appointments or commitments without AI-based prompts. Over time, this may lead to declining time-management skills and a passive approach to personal organization.

Choice-Making and Exploration

Streaming platforms and online stores use AI to recommend movies, books, songs, and products. While enhancing personalization, this narrows choices to algorithm-driven suggestions. Users may no longer browse or evaluate options critically, leading to intellectual homogeneity and reduced exposure to diverse or unfamiliar ideas (Pariser, 2011).

Social and Psychological Dimensions of Dependency

AI’s growing presence in decision-making also affects confidence and motivation. A person who always checks grammar with AI may feel incapable of writing unaided. A student who lets ChatGPT answer discussion questions may experience anxiety when asked to speak spontaneously. This psychological effect, known as learned helplessness, can set in even among high-performing individuals, gradually undermining confidence in their own intellect and judgment (Greenfield, 2015).

Summary table 4.0: example of dependency

Context	AI Tool/Use Case	Skill at Risk	Potential Consequences
Students	Essay generators, summarizers	Writing, research, critical thinking	Surface learning, plagiarism, disengagement
Educators	Grading software, content creators	Pedagogical insight, differentiation	Loss of nuance, standardization of teaching
Doctors	CDSS, diagnostic tools	Clinical reasoning	Diagnostic errors, loss of confidence
Lawyers	Legal research bots	Interpretation, argumentation	Overdependence, reduced analytical depth
GPS users	Navigation systems	Spatial awareness	Poor memory of routes, dependency on directions
Consumers	Content recommendation algorithms	Decision-making, critical analysis	Intellectual narrowing, reduced curiosity

Source: Author’s construction from the literature

THE CONSEQUENCES OF SKILL DEGRADATION

Academic Underperformance and Shallow Learning

One of the earliest signs of skill degradation caused by AI overuse is the emergence of shallow learning and subsequent academic underperformance. As students increasingly depend on AI tools to complete assignments, write essays, and generate study materials, they often bypass the deep cognitive processes essential for meaningful learning. While these tools may enhance productivity and improve presentation, they tend to encourage surface-level engagement (Kirschner & De Bruyckere, 2017). Students may appear proficient due to AI-assisted outputs, but when evaluated in contexts that demand independent thinking such as oral exams, case analysis, or in-class assessments, they often struggle to apply or transfer knowledge.

This discrepancy between appearance and actual competence creates a false sense of preparedness, where learners believe they understand a topic simply because they have interacted with AI-generated content. As a result, their academic performance becomes inconsistent: high in AI-supported tasks but weak in situations requiring unaided reasoning, retention, or synthesis. Over time, such dependency can undermine academic confidence, diminish intrinsic motivation, and weaken foundational skills that formal education aims to develop (Selwyn et al., 2020).

Decline in Creativity and Intellectual Risk-Taking

Another serious consequence of AI dependency is the decline in human creativity and a reduction in the willingness to take intellectual risks. AI systems are designed to generate responses based on established patterns and existing data. While this makes them reliable tools for routine tasks, it also means that their outputs are often conventional and predictable. When students or professionals rely on AI to brainstorm ideas, structure arguments, or design content, they may gradually stop exploring their own novel or unconventional ideas (Carr, 2020).

The presence of a readily available solution suppresses the natural discomfort that often drives creative insight. Instead of struggling with ambiguity or experimenting with different approaches, users are encouraged to settle for the most immediate and coherent AI-generated suggestion. This reliance not only dampens originality but also fosters intellectual conformity, as users replicate the same templates and thought structures produced by machines. In disciplines such as the arts, philosophy, and innovation-based industries, this trend is particularly alarming, as it undermines the very qualities – divergent thinking, persistence, and ideation – that fuel transformative contributions (Amabile, 1996).

Weakening of Professional Competence

The effects of AI-induced skill degradation are not confined to education; they are increasingly evident across various professional domains. In healthcare, for example, doctors are now supported by clinical decision-support systems that suggest diagnoses or treatment plans. While these tools can reduce errors and save time, they may also discourage practitioners from engaging in deep diagnostic reasoning. Over time, younger or less experienced doctors may become so accustomed to AI recommendations that their capacity for independent clinical judgment declines (Cabitza, Locoro, & Banfi, 2018).

Similar trends are visible in finance, where analysts rely on predictive models, and in legal professions, where AI tools summarize case law or generate contracts. Professionals who accept AI outputs without verification may lose touch with core disciplinary reasoning and gradually become deskilled. This dependence becomes particularly dangerous in novel, complex, or high-risk situations where human intuition, ethical discernment, and contextual understanding are indispensable. Moreover, the erosion of expertise diminishes job satisfaction and professional identity, as individuals feel less ownership of the work and more like passive supervisors of automated outputs (Susskind, 2019).

Loss of Metacognition and Self-Regulation

Metacognition, the ability to think about one’s own thinking, plays a critical role in effective learning and problem-solving. AI tools that offer instant answers, corrections, and summaries can inadvertently reduce the need for learners to reflect on their own understanding or evaluate their cognitive processes. When students rely on AI to highlight grammar issues, generate citations, or explain complex ideas, they are less likely to assess whether they truly grasp the material (Bjork & Bjork, 2020).

This pattern of dependency disrupts the development of self-regulatory skills such as goal-setting, strategic planning, monitoring progress, and adjusting learning approaches. Without regular practice in evaluating their own work, learners may fail to identify knowledge gaps or misunderstandings, leading to overconfidence and stagnation. In the long term, diminished metacognitive awareness limits students’ capacity to become independent, self-directed learners which is an essential attribute for success in academic, professional, and personal life (Zimmerman, 2002).

Psychological Dependency and Erosion of Confidence

As AI becomes a constant companion in cognitive tasks, many users begin to experience psychological dependency. Individuals who once approached writing, analysis, or planning with confidence may find themselves unable or unwilling to perform these tasks without AI assistance. This dependency often leads to learned helplessness, a psychological condition where individuals feel incapable of acting independently, even when they possess the required skills (Seligman, 1975; Greenfield, 2015).

The illusion of competence that AI fosters can collapse in moments of unassisted performance, leading to anxiety, frustration, and decreased self-esteem. Students may dread in-class tasks that require spontaneous thinking, while professionals may feel insecure when asked to present ideas or decisions developed without machine input. In both cases, the result is diminished confidence in one’s cognitive abilities and a growing fear of failure outside AI-supported environments (Bandura, 1997).

Equity Gaps and Digital Divide

AI dependency also contributes to a more complex form of digital inequality. Traditionally, digital divides referred to disparities in access to technology. Today, however, disparities are emerging not just in access, but in usage patterns and outcomes. Students with reliable access to AI tools may produce more polished work, yet lack deep comprehension. Conversely, students without AI may develop stronger foundational skills through manual effort, yet receive lower evaluations due to less refined presentation (Cukurova, Bennett, & Abrahams, 2018).

This paradox creates an invisible merit gap, where performance is judged more by output quality than by underlying competence. As AI tools become more integrated into education and employment, those who use them wisely and critically will likely advance, while those who misuse them or lack guidance may fall behind. Without deliberate policies and educational interventions, this divergence in AI literacy and dependency may deepen existing social and educational inequalities (Selwyn et al., 2020).

Devaluation of Human Expertise

Finally, as AI systems outperform humans in speed and consistency for many routine tasks, there is a growing risk of devaluing human judgment, experience, and insight. When AI-generated outputs are perceived as superior or sufficient, the unique contributions of human professionals such as empathy, ethical reasoning, contextual sensitivity, and creative problem-solving may be overlooked or undervalued.

In educational settings, the teacher’s role may shift from active facilitator to mere validator of AI-facilitated learning. In workplaces, human discretion may be sidelined in favor of algorithmic efficiency. This trend undermines professional dignity, erodes trust in human competence, and may discourage future generations from investing in skills and careers that appear replaceable. Moreover, it raises fundamental ethical concerns about how society values human intellect and what role it assigns to humans in an increasingly automated world (Luckin & Holmes, 2016; Russell & Norvig, 2021).

ETHICAL AND PEDAGOGICAL CONSIDERATIONS

The growing integration of AI into education and professional domains introduces not only cognitive and psychological concerns but also profound ethical and pedagogical challenges. As institutions and individuals adopt AI technologies to enhance productivity and streamline learning, they must grapple with critical questions related to fairness, transparency, academic integrity, and the evolving dynamics of teaching and learning. This section explores the ethical dilemmas and pedagogical tensions associated with AI overuse, particularly focusing on issues that influence educational equity, learner autonomy, and instructional responsibility.

One of the most pressing ethical concerns is the erosion of academic integrity. The ease with which AI can generate essays, solve complex problems, or write discussion posts has blurred the boundaries between original work and assisted output. Students may submit AI-generated content as their own, either knowingly or under the illusion that editing or paraphrasing is sufficient to claim authorship. This raises critical questions about plagiarism, authorship, and intellectual honesty. As AI tools become more fluent and difficult to detect, educational institutions must revise traditional definitions of cheating and develop new frameworks for evaluating student work (Selwyn et al., 2020).

Closely tied to this is the issue of fairness. AI use in education can unintentionally advantage certain students while disadvantaging others. Learners with better digital literacy or access to premium AI tools may produce more polished work, regardless of actual comprehension or effort. Meanwhile, students without these resources, or those who choose to work independently, may appear less competent. This creates an uneven playing field and challenges the principle of equal opportunity in assessment. Moreover, algorithms themselves may carry hidden biases, influencing what kind of feedback, content, or recommendations students receive. If left unchecked, these disparities can exacerbate educational inequities, particularly in under-resourced institutions or regions (Cukurova, Bennett, & Abrahams, 2018).

Transparency is another ethical imperative. Many AI tools function as “black boxes,” meaning users and educators do not fully understand how decisions are made, what data is used, or what biases are embedded in the algorithms. This lack of transparency is especially problematic in AI-driven grading systems, recommendation engines, or adaptive learning platforms. When students receive feedback or scores from AI without explanation, they are denied the opportunity to understand their performance or improve through reflection. Furthermore, instructors may inadvertently delegate their judgment to systems whose operations they do not fully comprehend, weakening their pedagogical role and diminishing accountability (Luckin & Holmes, 2016).

Data privacy and consent also present major ethical risks. AI platforms often collect vast amounts of user data, including writing samples, learning preferences, and behavioral patterns. In many cases, students are not fully aware of how their data is stored, processed, or shared. The use of such data, particularly by commercial AI vendors, raises concerns about surveillance, commodification of learning behavior, and long-term digital profiling. Ethical pedagogy requires that students be informed about how their information is used and be given the ability to opt out or restrict access. Institutions, in turn, must ensure that their adoption of AI complies with data protection regulations and is aligned with students’ rights to privacy and autonomy (Williamson & Eynon, 2020).

From a pedagogical perspective, the use of AI challenges traditional models of teaching and learning. Historically, education has emphasized effort, persistence, and the gradual development of understanding through struggle and iteration. AI, by offering instant solutions and feedback, may bypass this process. Learners accustomed to AI assistance may find it difficult to tolerate ambiguity, make mistakes, or engage in sustained problem-solving, all of which are essential to deep learning. This shift undermines the constructivist principles upon which much of modern education is based, where knowledge is actively constructed rather than passively received (Vygotsky, 1978; Bruner, 1966).

Moreover, AI alters the student-teacher dynamic. Teachers are no longer the sole sources of information, nor even the primary interpreters of knowledge. Instructors must now navigate classrooms where students use AI to challenge, verify, or even bypass their input. While this offers opportunities for more interactive, dialogic teaching, it also requires that educators be equipped to guide critical engagement with AI tools. Pedagogical strategies must evolve to help students reflect on when and how to use AI appropriately, and how to distinguish between meaningful learning and mere task completion (Holmes et al., 2022).

This raises the issue of faculty readiness. Many educators are still unfamiliar with AI technologies, let alone their ethical and pedagogical implications. Without adequate training and institutional support, they may feel ill-prepared to incorporate AI meaningfully into their teaching. This can lead to a defensive posture, in which instructors discourage or penalize AI use rather than channeling it toward productive learning. Effective integration requires a shift from resistance to strategic adaptation where AI is positioned not as a replacement for instruction, but as a complement to human mentorship and facilitation (Luckin & Cukurova, 2019).

Lastly, ethical pedagogy must consider the long-term impact of AI on learner identity and purpose. Education is not only about acquiring knowledge but also about becoming a particular kind of person – a critical thinker, a responsible citizen, a lifelong learner. When AI becomes overly dominant in educational processes, there is a risk that students will define success in terms of efficiency, correctness, or output quality, rather than curiosity, reflection, or engagement. Educators must therefore cultivate an environment in which AI use is contextualized, critically examined, and framed within broader humanistic goals of education.

FUTURE DIRECTIONS

As artificial intelligence continues to influence every facet of academic and professional life, addressing the risks of skill degradation becomes not just a pedagogical necessity but a strategic imperative. While the integration of AI into education and knowledge work is irreversible, its consequences, particularly the erosion of cognitive skills and autonomy, can be managed through intentional policies, institutional reform, and reflective practices. The future of AI in education should not aim to eliminate AI use, but rather to promote balanced, ethical, and empowering forms of integration that enhance rather than diminish human capabilities.

One of the most urgent future priorities is the development of clear institutional policies and guidelines on the responsible use of AI in academic settings. Universities and schools must move beyond ad hoc responses to student AI use and instead create structured frameworks that define acceptable practices, academic boundaries, and pedagogical goals. These policies should distinguish between permissible assistance (e.g., grammar checking or citation formatting) and prohibited use (e.g., full essay generation or automated exam completion). Moreover, academic integrity policies must evolve to incorporate AI-specific language, helping both students and faculty navigate ethical gray areas without fear or ambiguity (Holmes et al., 2022).

In tandem with policy development, institutions must invest in comprehensive faculty training and professional development. Many instructors remain unfamiliar with the capabilities, limitations, and risks of AI tools. Future-ready educational institutions must equip teachers not only to detect AI misuse but to engage with it pedagogically. This includes training on how to design AI-resilient assessments, use AI to personalize instruction, and guide students toward critical, intentional use of these tools. Faculty must be supported in shifting from gatekeepers of knowledge to facilitators of AI-enhanced learning environments (Luckin & Cukurova, 2019).

For students, the future calls for robust AI literacy education, embedded into curricula across disciplines. Just as digital literacy became a key educational goal in the early 21st century, AI literacy will be essential for the next generation of learners. Students must be taught not only how to use AI tools, but when and why to use them. This involves understanding how algorithms work, recognizing the limitations of machine-generated knowledge, and reflecting on the cognitive trade-offs of automation. Such literacy programs should also include ethical reasoning, helping learners consider the societal and personal implications of relying on intelligent systems (Williamson & Eynon, 2020).

Assessment reform is another critical area for future action. Traditional assessment models, which emphasize product over process, are vulnerable to AI manipulation and do little to foster independent thinking. Educational institutions must explore alternative forms of evaluation that emphasize creativity, problem-solving, and reflective learning. This could include project-based assessments, oral examinations, collaborative tasks, and portfolios that capture the evolution of a learner’s thought process over time. By valuing the journey of learning as much as its outcomes, these assessments can better reveal genuine competence and reduce the temptation to outsource thinking to machines (Selwyn et al., 2020).

Furthermore, the future of AI integration must be guided by human-centered design principles. AI tools used in education should be designed not only for efficiency but also for learner development. Designers, developers, and institutions should collaborate to create systems that scaffold rather than replace human effort – tools that prompt reflection, offer feedback loops, and encourage metacognitive engagement. For example, AI-based writing assistants could provide reasoning explanations for suggestions, or learning platforms could ask users to justify their prompts or evaluate AI outputs. Such features would help preserve critical thinking while still leveraging AI’s strengths (Luckin et al., 2016).

In addition to design, there is a pressing need for ongoing interdisciplinary research on the long-term effects of AI dependency. While the literature on AI in education is growing, more empirical studies are needed to understand how overreliance affects skill acquisition, retention, and transfer across different age groups, disciplines, and learning contexts. Research should also explore cultural and regional variations in AI use, especially in under-resourced or marginalized communities. These insights can inform more inclusive and equitable AI policies that do not reinforce existing educational disparities (Cukurova et al., 2018).

Finally, the future direction must include a renewed emphasis on the philosophical and ethical foundations of education. As AI becomes more embedded in learning environments, institutions must ask: What does it mean to be an educated person in the age of intelligent machines? What skills, dispositions, and values do we wish to cultivate in learners? How can we ensure that AI supports rather than supplants these humanistic goals? These questions are not merely technical, they are existential, and answering them will require the collaboration of educators, philosophers, technologists, and policymakers alike (Selwyn et al., 2020; Bruner, 1966).

CONCLUSION

The rapid adoption of artificial intelligence in higher education has brought about both unprecedented opportunities and unforeseen challenges. While AI has demonstrated immense potential to enhance learning efficiency, personalize educational experiences, and support instructional delivery, its uncritical and widespread use has also led to the silent erosion of key human capabilities. This paper has examined the phenomenon of AI-induced skill degradation, particularly highlighting how students and professionals may gradually lose essential competencies such as critical thinking, creativity, metacognition, and autonomy – through overdependence on intelligent systems.

At the heart of this degradation lies what has been termed the illusion of competence, a misleading sense of mastery fostered by AI’s fluency and convenience. Learners who habitually rely on AI-generated content may appear productive and proficient, yet often lack the deeper understanding and adaptive skills required in real-world contexts. The same pattern holds true in professional environments, where automation, while efficient, can displace judgment, reduce intellectual engagement, and ultimately deskill human practitioners.

The consequences of this phenomenon are wide-ranging. In academic settings, shallow learning, academic dishonesty, and inequitable assessment outcomes threaten the integrity of education. In the workplace, reduced expertise, misplaced trust in automation, and diminished confidence in human judgment weaken professional identity and performance. Psychologically, AI dependency fosters learned helplessness and anxiety, eroding users’ belief in their own abilities. Societally, it risks amplifying inequality, narrowing intellectual diversity, and devaluing the uniquely human contributions of reason, ethics, and creativity.

To address these issues, the future of AI in education must be guided by ethical reflection and pedagogical intentionality. Institutions should develop clear policies and frameworks that define responsible AI use, support faculty in redesigning curriculum and assessment, and embed AI literacy into learning objectives. Designers must create tools that scaffold rather than replace human effort, while educators must cultivate learning environments that foster critical engagement with AI rather than passive consumption. Above all, stakeholders must reaffirm education’s broader purpose, not merely the efficient transfer of information, but the cultivation of thoughtful, autonomous, and ethically grounded individuals.

Artificial intelligence is not inherently detrimental to human learning. It becomes so when its use is uncritical, unregulated, and untethered from pedagogical and ethical foundations. The challenge ahead is not to resist AI, but to humanize it to ensure that, as machines grow smarter, we become not more dependent, but more discerning, resilient, and empowered in our learning and professional lives. Only then can AI serve as a tool that extends human potential rather than one that replaces it.

REFERENCES

1. Amabile, T. M. (1996). Creativity in context. Westview Press.
2. Bandura, A. (1997). Self-efficacy: The exercise of control. W.H. Freeman.
3. Bjork, E. L., & Bjork, R. A. (2020). Desirable difficulties in theory and practice. Journal of Applied Research in Memory and Cognition, 9(4), 475–479. https://doi.org/10.1016/j.jarmac.2020.09.003
4. Bloom, B. S. (1956). Taxonomy of educational objectives: The classification of educational goals. Longmans.
5. Bruner, J. S. (1966). Toward a theory of instruction. Harvard University Press.
6. Cabitza, F., Locoro, A., & Banfi, G. (2018). Machine learning in medicine: A survey of current clinical applications. Journal of Biomedical Informatics, 86, 34–54. https://doi.org/10.1016/j.jbi.2018.08.007
7. Carr, N. (2020). The shallows: What the Internet is doing to our brains (10th anniversary ed.). W.W. Norton & Company.
8. Cukurova, M., Bennett, J., & Abrahams, I. (2018). Students’ knowledge acquisition and ability to apply knowledge into different science contexts in two different independent learning settings. Research in Science & Technological Education, 36(1), 17–34. https://doi.org/10.1080/02635143.2017.1310092
9. Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive–developmental inquiry. American Psychologist, 34(10), 906–911. https://doi.org/10.1037/0003-066X.34.10.906
10. Greenfield, S. (2015). Mind change: How digital technologies are leaving their mark on our brains. Random House.
11. Holmes, W., Bialik, M., & Fadel, C. (2022). Artificial intelligence in education: Promises and implications for teaching and learning. Center for Curriculum Redesign.
12. Kirschner, P. A., & De Bruyckere, P. (2017). The myths of the digital native and the multitasker. Teaching and Teacher Education, 67, 135–142. https://doi.org/10.1016/j.tate.2017.06.001
13. Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016). Intelligence unleashed: An argument for AI in education. Pearson Education. https://dera.ioe.ac.uk/27802/
14. Luckin, R., & Cukurova, M. (2019). Designing educational technologies in the age of AI. Nature Human Behaviour, 3(3), 250–252. https://doi.org/10.1038/s41562-019-0535-0
15. Nilsson, N. J. (2019). The quest for artificial intelligence: A history of ideas and achievements. Cambridge University Press.
16. Russell, S. J., & Norvig, P. (2021). Artificial intelligence: A modern approach (4th ed.). Pearson Education.
17. Selwyn, N., Hillman, T., Eynon, R., Ferreira, G., & Knox, J. (2020). What’s next for Ed-Tech? Critical hopes and concerns for the 2020s. Learning, Media and Technology, 45(1), 1–6. https://doi.org/10.1080/17439884.2020.1694945
18. Seligman, M. E. P. (1975). Helplessness: On depression, development, and death. W.H. Freeman.
19. Sparrow, B., Liu, J., & Wegner, D. M. (2011). Google effects on memory: Cognitive consequences of having information at our fingertips. Science, 333(6043), 776–778. https://doi.org/10.1126/science.1207745
20. Susskind, R. (2019). Tomorrow’s lawyers: An introduction to your future (2nd ed.). Oxford University Press.
21. Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.
22. Williamson, B., & Eynon, R. (2020). Historical threads, missing links, and future directions in AI in education. Learning, Media and Technology, 45(3), 223–235. https://doi.org/10.1080/17439884.2020.1798995
23. Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory Into Practice, 41(2), 64–70. https://doi.org/10.1207/s15430421tip4102_2