Applied Psychology and Cognitive Science

Artificial Intelligence in Psychological Practice: A Systematic Review of Opportunities, Challenges, and Ethical Considerations

Rizky Aditya Pranawa — 2026-10-26

Artificial intelligence (AI) is increasingly reshaping psychological practice by extending the capabilities of clinical assessment, personalized intervention, decision support, and continuous behavioral monitoring. This systematic review critically synthesized current evidence on the clinical opportunities, implementation barriers, and ethical implications associated with the integration of AI into psychological care. Contemporary applications encompassed machine learning models for diagnostic screening and suicide risk prediction, conversational agents delivering cognitive behavioral therapy, clinician decision support systems, and digital phenotyping through smartphones and wearable sensors. Evidence from randomized controlled trials and systematic reviews suggested that AI assisted interventions could enhance accessibility, scalability, patient engagement, and outcomes related to depressive symptoms. However, evidence regarding anxiety, stress, and other psychological outcomes remained heterogeneous, limiting definitive conclusions regarding clinical effectiveness. Translation into routine psychological practice was further constrained by small and insufficiently representative datasets, limited external and prospective validation, algorithmic bias, inadequate model interpretability, uncertain clinical accountability, workforce readiness, and concerns regarding potential disruption of therapeutic relationships. Ethical and governance challenges were particularly prominent and included privacy, data security, informed consent, transparency, accountability, equitable access, and compliance with rapidly evolving regulatory frameworks. Collectively, the evidence indicated a substantial gap between technological capability and clinically validated implementation. Future research should prioritize representative and culturally diverse datasets, prospective multicenter validation, longitudinal evaluation of clinical outcomes, systematic assessment of algorithmic bias, and interdisciplinary collaboration among psychologists, AI developers, healthcare institutions, policymakers, and regulators. Rather than functioning as a substitute for psychological expertise, AI should be positioned as an augmentative clinical tool operating within clearly defined professional and ethical boundaries. Responsible integration therefore requires a human centered framework in which clinical judgment, therapeutic relationships, patient autonomy, transparent governance, and robust ethical safeguards remain central to psychological care.

Factors Influencing Public Trust in AI Decision Support Systems: A Cognitive and Behavioral Perspective

Dimas Ardhana Prasetyo — 2026-10-26

Public trust represents a critical prerequisite for the responsible adoption of artificial intelligence based decision support systems (AI DSS), particularly in high stakes contexts where algorithmic recommendations may substantially influence consequential human decisions. Despite increasing reliance on AI DSS, the cognitive and behavioral mechanisms underlying public trust remain insufficiently understood, particularly regarding how transparency interacts with individual resistance to algorithmic decision making. This study examined the cognitive and behavioral determinants of trust in AI DSS using a cross sectional survey of 412 adults. Perceived transparency, algorithm aversion, anthropomorphism, need for cognition, general attitudes toward AI, perceived risk, and prior AI experience were examined as potential determinants of trust. Hierarchical multiple regression and moderation analyses were conducted to identify significant predictors and conditional relationships, while analysis of variance was used to assess differences in trust according to prior AI experience. The cognitive and behavioral factors collectively explained approximately 50% of the variance in trust. Perceived transparency emerged as the strongest positive predictor, followed by favorable attitudes toward AI, whereas perceived risk and algorithm aversion significantly reduced trust. Importantly, algorithm aversion significantly moderated the relationship between transparency and trust, such that the positive effect of transparency weakened as algorithm aversion increased. Trust also differed significantly according to prior AI experience, with highly experienced participants reporting greater trust than those with limited experience. These findings demonstrate that transparency, although important, is insufficient as a standalone mechanism for establishing trust in AI DSS. The study advances understanding of AI trust by demonstrating that the effectiveness of transparency depends partly on users’ preexisting cognitive and behavioral dispositions toward algorithmic systems. Consequently, responsible trust building strategies should integrate understandable and context appropriate explanations, effective risk communication, meaningful user experience, and explicit consideration of individual differences in attitudes toward algorithmic decision making.

Psychological Predictors of Employee Adaptation to Artificial Intelligence in the Workplace

Tharushi Sewwandi Rathnayake — 2026-10-26

Organizations are increasingly integrating artificial intelligence (AI) into routine work processes, fundamentally reshaping how employees perform tasks, make decisions, and interact with emerging technologies. However, employees differ considerably in their capacity to adapt to AI driven workplace transformation, highlighting the importance of understanding the psychological mechanisms underlying successful adaptation. This study investigated the psychological determinants of employee adaptation to AI by integrating AI self efficacy, AI anxiety, psychological capital, perceived organizational support, and openness to experience within a unified theoretical framework. A cross sectional survey design was employed, and hierarchical multiple regression and bootstrapped mediation analyses were conducted to examine the proposed relationships. The results demonstrated that AI self efficacy, psychological capital, and perceived organizational support positively predicted employee adaptation to AI, whereas AI anxiety negatively predicted adaptation. Psychological capital further partially mediated the relationship between AI anxiety and employee adaptation, indicating that positive psychological resources may reduce, although not eliminate, the adverse influence of AI related anxiety. These findings extend the Job Demands Resources framework to AI driven organizational transformation by demonstrating how technology related demands and personal and organizational resources jointly contribute to employee adaptation. The study further highlights psychological capital as an important mechanism connecting AI related anxiety with adaptive workplace outcomes. From a practical perspective, organizations should complement technical AI training with interventions that strengthen employees’ confidence, resilience, optimism, psychological readiness, and perceptions of organizational support. Such an integrated approach may facilitate more sustainable employee adaptation and reduce psychological barriers during AI implementation.

Cyberpsychology in the Era of Generative AI: Human Behavior, Trust, and Online Decision-Making

Amira N. Kartika — 2026-10-26

The rapid diffusion of generative artificial intelligence (AI) into everyday digital environments has transformed the ways individuals communicate, seek information, form relationships, and make decisions online. This narrative review synthesizes current cyberpsychological research on human behavior, trust, and online decision-making in the context of generative AI systems such as large language model chatbots, conversational agents, and synthetic media generators. Drawing on theoretical frameworks including the Computers Are Social Actors paradigm, privacy calculus theory, and algorithm aversion/appreciation models, the review examines three interrelated domains: (a) psychological responses to generative AI, including anthropomorphism, self-disclosure, cognitive offloading, and emotional dependence; (b) mechanisms of trust formation, calibration, and erosion in human-AI interaction; and (c) the behavioral and epistemic consequences of generative AI for online decision-making, including algorithm reliance, misinformation susceptibility, and privacy trade-offs. The review integrates evidence from 25 peer-reviewed journal articles and presents four synthesis tables and a conceptual radial diagram mapping the psychological pathways linking generative AI use to trust and decision outcomes. Findings indicate that trust in generative AI is shaped jointly by system-level cues (transparency, anthropomorphic design) and user-level dispositions (privacy concern, prior experience), while overreliance on generative AI is associated with reduced critical thinking and, in some populations, increased loneliness. Implications for digital literacy, platform design, and future cyberpsychological research are discussed.

Cross-Cultural Perspectives on Psychological Resilience: A Review

Meera S. Rajagopal — 2026-10-26

Psychological resilience, broadly defined as positive adaptation in the context of significant adversity, has become a prominent construct in developmental and clinical psychology. However, foundational theories and measures of resilience have largely been developed within Western, individualistic cultural contexts, raising concerns about their cross-cultural applicability. This review synthesizes evidence on how cultural contexts shape the conceptualization, expression, and measurement of psychological resilience. Eighteen peer-reviewed journal articles were reviewed and synthesized across four interconnected domains: theoretical and conceptual foundations of resilience, particularly the transition from trait-based perspectives to social-ecological models; methodological approaches to cross-cultural resilience research, including standardized scale validation and mixed-methods measure development; cultural variation in resilience processes among collectivist, Indigenous, and refugee populations; and developmental and lifespan considerations in cross-cultural resilience research. Four synthesis tables and a conceptual radial framework were used to integrate and map the major themes identified across the literature. The findings demonstrate that resilience is better understood as a culturally situated, socially embedded, and ecologically influenced process rather than as a culturally invariant individual trait. Culturally grounded measurement approaches, including the Child and Youth Resilience Measure, appear particularly useful for capturing contextually meaningful individual, relational, community, and cultural resources that may be overlooked by direct translations of measures developed in Western settings. These findings highlight the importance of culturally responsive conceptualization and assessment of resilience and provide implications for cross-cultural measurement, intervention development, and future research.