This critical narrative review synthesizes evidence from the past decade on AI dependency among college students, focusing on conceptualizations, theoretical frameworks, prevalence and demographic variations, measurement tools, determinants, intervention strategies, and research gaps.
Abstract
The rapid diffusion of generative artificial intelligence (GenAI), particularly large language model (LLM) applications, has fundamentally reshaped university students’ learning, problem-solving, and socio-emotional practices. Alongside clear benefits in efficiency and personalized support, concerns are emerging about AI dependency, understood here as an emerging, non-diagnostic educational-psychological construct rather than a formal clinical disorder. This critical narrative review synthesizes evidence from the past decade on AI dependency among college students, focusing on conceptualizations, theoretical frameworks, prevalence and demographic variations, measurement tools, determinants, intervention strategies, and research gaps. Across studies, an “ask-AI-first” pattern appears increasingly common, although frequency of use should not be equated with maladaptive dependency. The central concern is not AI use itself but patterns of cognitive delegation, emotional reliance, and reduced self-regulation that may be associated with weaker independent engagement, academic-integrity risks, and distress when AI is unavailable. The literature also reveals substantial fragmentation in assessment, although several emerging tools demonstrate promising psychometric properties. Determinants appear multi-level, involving individual vulnerabilities, academic pressure, contextual norms, and technological affordances. Current recommendations emphasize promoting augmentation over automation, strengthening AI literacy and critical evaluation skills, redesigning assessments to preserve cognitive engagement, and implementing supportive interventions for at-risk students. Future research should prioritize longitudinal, cross-cultural, and multi-method designs to clarify developmental trajectories, thresholds for maladaptation, and evidence-based intervention pathways.
Educational institutions have adopted Artificial Intelligence (AI) tools at a pace that has outstripped careful study of their cognitive consequences for students. This paper synthesizes findings from 25 recent journal articles and empirical studies to examine how AI integration in school, undergraduate, and graduate settings shapes critical thinking, logical reasoning, and broader cognitive development. Generative large language models, intelligent tutoring systems, and adaptive learning platforms clearly strengthen personalized instruction and raise measurable academic performance; the evidence reviewed here, however, points to a separate risk: when such tools are adopted without safeguards, they can gradually erode independent reasoning, deductive problem-solving, and metacognitive self-regulation. One especially striking pattern emerges from the empirical literature: students who rely heavily on AI support tend to earn higher grades yet perform worse on assessments that require unassisted thinking, a divergence they term the grade-competence gap. To address this tension, the paper proposes the CLEAR Framework: contextual use, literacy, engagement, assessment reform, and reflection, a five-component approach for capturing AI's instructional value without undermining students' cognitive growth. The paper closes with practical recommendations for educators, curriculum designers, and policymakers navigating AI's growing role in education.
Ashok Sadavare, D. Kumbhar, Aditya Sadavare· Journal of Data Engineering...· 0 citations
The rapid proliferation of artificial intelligence (AI) technologies—including machine learning
algorithms, natural language processing tools, and AI-powered chatbots—has begun to reshape
the landscape of psychological practice, raising both opportunities and concerns for the
profession. This paper presents a secondary, desk-based review synthesizing existing literature on
the integration of AI into psychological assessment, diagnosis, intervention, and administrative
practice. Drawing on peer-reviewed journal articles, professional guidelines, and policy
documents published between 2018 and 2025, the review examines three central themes: the
emerging roles AI tools play in augmenting psychologists’ clinical and research functions; the
competencies psychologists require to ethically and effectively engage with AI-driven
technologies; and the conceptual, ethical, and practical challenges associated with their adoption,
including concerns regarding data privacy, algorithmic bias, the therapeutic alliance, and
professional accountability. Findings suggest that while AI holds considerable promise for
enhancing diagnostic accuracy, expanding access to mental health services, and reducing
clinician workload, its integration is hindered by gaps in psychologists’ digital literacy,
inconsistent regulatory frameworks, and unresolved questions about the preservation of human
centered care. The review concludes by identifying priority areas for training curricula, ethical
guideline development, and future empirical research. This synthesis contributes to ongoing
professional discourse by consolidating fragmented evidence into a coherent framework for
understanding AI’s evolving role in psychological practice.
Michael Mbiriri· Journal of Humanities and So...· 0 citations
Generative artificial intelligence (GenAI) is rapidly changing professional language education, yet the behavioral and cognitive mechanisms through which learners engage with GenAI-mediated language tasks remain insufficiently synthesized. This critical integrative review reframes translation and interpreting education as a high-cognitive-load case of professional language learning in which students must evaluate AI output, regulate feedback use, and remain accountable for meaning across languages. A documented evidence-selection workflow was applied to an author-curated bibliographic corpus and a supplementary source-expansion corpus comprising 690 potentially relevant records, from which 21 core studies were selected for focused synthesis. The included studies directly addressed GenAI, translation/interpreting education, AI-generated feedback, post-editing, learner revision, interpreter assessment, or AI literacy. The synthesis is theoretically grounded in cognitive load theory, self-regulated learning, feedback literacy, trust in automation, and social-cognitive accounts of agency. Across the reviewed evidence, learners’ engagement with GenAI involves cognitive load redistribution, trust calibration, feedback uptake, metacognitive monitoring, affective responses, and behavioral revision decisions. The review proposes a behavioral model that links technological mediation, cognitive appraisal, self-regulated behavioral engagement, pedagogical regulation, observable process evidence, and professional agency. We argue that the central educational challenge is not whether GenAI improves single-task language performance, but how learners develop calibrated trust, critical judgment, self-regulated feedback use, and professional agency in human–AI language-learning environments. The framework may also inform adjacent AI-mediated language-learning contexts where learners must evaluate feedback, revise language, and justify communicative choices.
The integration of Artificial Intelligence (AI) in language education has redefined pedagogical practices, particularly in English as a Foreign Language (EFL) learning. While AI-powered tools promise to enhance personalization, engagement, and learner-centered approaches, their impact on learners' autonomy remains a subject of debate. This study aims to explores the dual role of AI in fostering, or potentially hindering, the development of autonomy among Moroccan university EFL learners. Drawing on both qualitative and quantitative data, the study investigates how AI tools influence learners’ decision-making, self-regulation, and motivation in language learning contexts. Combining quantitative data from questionnaires distributed to 200 EFL students with qualitative insights from interviews conducted with 10 professors, the research assesses how AI affects students' capacity for self-directed learning. The findings reveal prevalent use of AI-powered platforms such as ChatGPT, Google Translate, Grammarly, and Duolingo among EFL learners, primarily for language practice, grammar correction, vocabulary acquisition, translation tasks, and exam preparation. Both students and professors acknowledged that these tools can significantly enhance language learning autonomy by enabling personalized, flexible learning experiences and fostering increased responsibility and motivation. However, concerns were also raised about the potential downsides, particularly overreliance on AI, reduced engagement in critical thinking, and threats to academic integrity. While some professors observed improved student engagement and self-directed learning, others noted a decline in cognitive effort when learners relied too heavily on AI-generated content. Overall, AI emerged as a double-edged sword; it offers valuable support for autonomous learning, yet its unchecked use may hinder essential skill development. The study concludes that the thoughtful and guided integration of AI is crucial to ensure it strengthens rather than weakens EFL learner autonomy.
Hanae Ait Hattani· Proceedings of The World Con...· 0 citations
The use of large language models (LLMs) and generative artificial intelligence (AI) tools (e.g., ChatGPT, AI-assisted supervision platforms, and AI-powered role-play simulations) in counselor education is changing the way counselors-in-training (CITs) are trained and has both risks and benefits for the field. This conceptual article applied Bronfenbrenner’s ecological systems theory to evaluate the impact of these technologies at multiple levels within the counseling training environment. The microsystem is described as the interactions between CITs and the AI they use to support their training experiences (e.g., AI-based learning tools, role-playing simulations). The mesosystem refers to the movement of AI across different training environments (e.g., classroom, clinical supervision, and field work), with the potential for cohesive communication or disjointed messages. Exosystem factors, such as institutional policies and procedures, accreditation requirements and standards, and resource inequities, create the context in which AI can be integrated into counselor education programs. Macrosystem factors include the broader societal values, professional ethics, and commitment to the values of social justice and cultural humility that may converge with the integration of AI into counselor education. The chronosystem is concerned with time-related issues; specifically, the developmental stages of CITs and the evolving pace of new AI technology. Utilizing an ecological framework, counselor educators can take a proactive approach to address the systematic efforts of AI and support the incorporation of critical AI literacy into all aspects of counselor education (curriculum design, supervisory processes, policy development, and research agendas). Ultimately, the decisions made now by counselor educators will determine whether AI enhances or diminishes the humanistic, multicultural, and social justice values that are fundamental to the counseling profession.
Lauren Dollarhide, Leo Gonzalez, Rachel M. Smith· Journal of Technology in Cou...· 0 citations
Large language models (LLMs) are increasingly integrated into students' learning processes, serving not only as cognitive tools for academic support but also as conversational partners that influence learners' emotional experience. However, existing LLM-based learning systems are not explicitly designed to account for learners' emotional vulnerability, and institutions often deploy such models as black-box services, limiting the feasibility of model-level modification. In this work, we investigate the prevalence of emotionally risky responses in realworld LLM-assisted learning and propose a non-invasive, posthoc emotion-value gated framework that enhances emotional safety and supportive communication without altering underlying model architectures. Through a large-scale analysis of 42,172 authentic student-AI interactions and a controlled deployment involving 24 students, we show that emotionally risky responses occur with meaningful frequency in baseline usage and can be substantially reduced by our approach. Expert evaluations demonstrate significant improvements in emotional safety, supportiveness, and overall response quality, while student studies indicate increased comfort, engagement, and willingness to use the emotion-enhanced AI, with no perceived loss in response correctness. These findings highlight the importance of emotional governance in educational AI and demonstrate that system-level, deployable interventions can meaningfully improve students' learning experience in LLM-assisted environments.
Yicheng Sun, J. Keung, H. Yu et al.· Annual International Compute...· 0 citations
