ABSTRACT The growing presence of agentic AI in public governance is transforming the once familiar, human-bound principal–agent relationship into a multi-directional socio-technical challenge. This article provides a structured analysis of these challenges by dissecting how AI-related agency problems manifest across three critical delegation interfaces: trust delegation between human users (principals) and AI agent systems; hierarchical delegation between senior officials (principals) and lower-level officials (agents) augmented by AI agents; and contractual delegation where government officials (principals) outsource AI development to private firms (agents). At each interface, the authors clarify how delegation drift may arise due to the intensified information asymmetry and goal misalignment, while creating new accountability gaps. Building on this analysis, the article concludes with a discussion on governing AI agents and proposes a forward-looking research agenda for public administration scholars.

Psychosocial interventions support functional recovery and social integration for people with severe mental illness (SMI); however, their implementation remains fragmented and inconsistent. This rapid scoping review maps the landscape of psychosocial interventions for adults with SMI in high-income countries, identifying their key components, implementation challenges, and strategies to optimise delivery.Following the Cochrane rapid review guidelines, we searched MEDLINE and CINAHL (January 2009 - May 2024) and conducted manual reference screening. Eligible studies focused on non-pharmacological, community-based interventions. Thematic analysis was used to identify implementation barriers and enablers.Of 8,624 screened records, 464 studies met inclusion criteria, covering 15 intervention types. Cognitive-focused interventions (n = 100) supported employment (n = 86), and behavioural therapies (n = 66) were most studied. Outcomes focused on symptom (n = 179), daily living skills (n = 160), and cognitive states (n = 157). Implementation was often constrained by systemic, organisational, and individual-level barriers. Systemic challenges included fragmented services, stigma, cultural and linguistic barriers, and economic constraints, highlighting the need for integrated care models, policy reforms, and culturally responsive approaches. Organisational challenges such as staff resistance, insufficient training, and resource limitations underscored the importance of leadership, stakeholder engagement, and investment. At the individual level, low motivation, logistical difficulties, trauma histories, and goal misalignment reduced engagement and retention, highlighting the need for flexible, person-centred, trauma-informed approaches, strong social networks, and a balance between structure and adaptability.Selecting, adapting, and funding psychosocial interventions remain complex. This review provides a foundation for future systematic reviews of homogeneous intervention subsets to better inform policy and practice.

ObjectiveTraumatic stress among families of critically ill patients is common and associated with interpersonal conflict between families and intensive care unit (ICU) clinicians. This qualitative study aimed to characterize clinician documentation of family members with high levels of traumatic stress in the electronic health record (EHR).MethodsMechanically ventilated patients with surrogate decision makers reporting a Post-Traumatic Stress Symptoms Scale score >35 (consistent with post-traumatic stress disorder) were included. All electronic notes from patients' ICU stays were included. Thematic analysis used a structured codebook focused on surrogate behaviors, relationships, and coping strategies; resources provided to surrogates; and medical decision making.ResultsTwenty-five surrogates were largely female (n = 22, 88.0%) and had a median age of 48 years (interquartile range [IQR] 41-55 years) and median PTSS-10 score of 46 (IQR 38-53). Most of their critically ill loved ones (n = 17, 68%) survived the hospitalization. Three major themes emerged: (1) documentation of surrogate traumatic stress was inconsistent unless it disrupted clinical care; (2) surrogate-clinician conflict and goal misalignment intensified surrogate distress; and (3) multidisciplinary team members were crucial for providing emotional support and bridging communication.ConclusionThese findings suggest the need for standardized approaches to identify and address surrogate traumatic stress in the ICU. Investment in multidisciplinary teams is also essential-not only to ease surrogate distress but also to facilitate trust, communication, and collaborative decision-making. Investing in these resources and bringing awareness to how our documentation can perpetuate stigma are key steps toward reducing re-traumatization and advancing person-centered care in the ICU.

Abstract As workplace inequality persists, understanding the psychological processes motivating collective action engagement becomes imperative. In one domain of collective action, labor unionization, membership in the United States has continuously declined since the 1980s. Despite this decline, recent surveys suggest that pro‐union sentiment is rising. Amid workplace inequity and rising public interest in unionization, we draw from research on collective action and organizational identification to examine whether exposure to worker‐organization conflict and information about unions changes how workers evaluate existing and novel organizational grievance procedures, and consequently, increases pro‐union sentiment. Across four studies ( N = 3,143), we find that exposure to workplace conflicts and information about unions reduces workers’ perceived alignment with organizational grievance procedures and increases pro‐union attitudes, relative to control conditions where participants are not exposed to information about unions. These findings have implications for employee wellbeing and for the psychological processes implicated in collective action organizing.

Artificial Intelligence (AI) applications in healthcare have significant potential to address the unmet needs of older adults. To successfully adopt and implement AI in the care of older adults, it is critical to understand stakeholders' perspectives. We sought to explore the perceived benefits and concerns among stakeholders about AI applications in caring for older adults. We conducted individual semi-structured interviews with five groups of stakeholders: older adults and caregivers, clinicians, health system and health insurance plan leaders (payers), investors, and technology developers. Interviews asked about the perceived role of AI in the care of older adults, the perceived benefits and concerns regarding AI, and suggestions for mitigating the concerns. Interviews were audio recorded and transcribed verbatim. We used thematic content analysis to code the transcripts. Overall, 49 participants completed interviews: older adults/caregivers (n = 15), clinicians (n = 15), payers (n = 8), investors (n = 5), and technology developers (n = 6). We identified three themes. (1). Stakeholders reported multiple benefits of AI and identified several roles for its use in the care of older adults. (2). Stakeholders expressed concerns about AI, including worsening social isolation, high cost, propagating ageism, goal misalignment, and scams/misuse of AI; views on privacy concerns were mixed. (3). Stakeholders suggested potential solutions, such as setting appropriate guardrails, to mitigate concerns about AI. Given the complexity and significant unmet needs among older adults, AI's potential benefits and harms are both heightened in this population. Appropriate guardrails are needed to leverage the benefits of AI while mitigating potential harms. Our findings have implications for technology developers to design innovations that align with the stakeholders' perceived roles for AI, for regulatory bodies to incorporate stakeholders' concerns when developing AI regulations, and for health systems and end-users of technology to critically evaluate a product regarding its affordability and impact on social isolation and ageism.

ABSTRACT The rapid adoption of artificial intelligence (AI) is shifting from tools that assist human tasks toward self‐directed, agentic AI systems capable of planning and executing complex goals with minimal oversight. However, a clear understanding of what distinguishes these systems from conventional AI agents and generative AI is lacking, obscuring their unique opportunities and risks. To this end, this article addresses that gap by defining the core concepts, technologies, and management approaches for agentic AI systems, which utilize planning, shared memory, tools, and multi‐agent teamwork to complete complex tasks autonomously. By contrasting this paradigm with its predecessors, the paper synthesizes recent technical surveys, governance proposals, and early industrial deployments to highlight that while agentic AI enables transformative applications like end‐to‐end process automation and adaptive decision support, it also introduces significant challenges, including cascading errors, goal misalignment, and regulatory gaps. Finally, this paper concludes with strategic guidance for organizations and consumers to adopt the capabilities of these systems responsibly, emphasizing the imperative of maintaining transparency, accountability, and human oversight.

This study used semi-structured interviews with 11 key opinion leaders to explore why dermatologist involvement in innovation remains limited despite growing investment in the field. It found that barriers such as limited early exposure to industry, time constraints, and goal misalignment hinder engagement, but also highlighted solutions like integrating industry exposure into training and fostering new funding models. These findings suggest actionable strategies to bridge gaps between clinicians and industry and promote dermatologic innovation.

Although the detrimental effects of exploitative leadership have been widely revealed, scholarly understanding of its formation mechanisms remains insufficient. Addressing this gap, the present study draws upon trait activation theory to investigate the antecedents and formation mechanisms of exploitative leadership through the dual lenses of personality traits and organizational contexts. Through an empirical examination of questionnaire data from 422 leader–subordinate dyads in Chinese organizations, the results of this study reveal that: three leader personality traits—self-interest, other-oriented perfectionism, and high power distance orientation—significantly and positively predict exploitative leadership behaviors; three organizational contexts—task challenge, flexible status conferral, and leader–subordinate goal misalignment—likewise serve as positive predictors of exploitative leadership behaviors; through trait activation effects, flexible status conferral significantly strengthens the positive relationships between all three personality traits and exploitative leadership, while task challenge and goal misalignment specifically amplify the positive relationships of exploitative leadership with high power distance orientation and self-interested personality, respectively. The research findings are of great value for improving the theoretical framework of exploitative leadership and guiding management practice.

Agentic AI and Large Language Models (LLMs) are transforming how language is understood and generated while reshaping decision-making, automation, and research practices. LLMs provide underlying reasoning capabilities, and Agentic AI systems use them to perform tasks through interactions with external tools, services, and Application Programming Interfaces (APIs). Based on a structured scoping review and thematic analysis, this study identifies that core challenges of LLMs, relating to security, privacy and trust, misinformation, misuse and bias, energy consumption, transparency and explainability, and value alignment, can propagate into Agentic AI. Beyond these inherited concerns, Agentic AI introduces new challenges, including context management, security, privacy and trust, goal misalignment, opaque decision-making, limited human oversight, multi-agent coordination, ethical and legal accountability, and long-term safety. We analyse the applications of Agentic AI powered by LLMs across six domains: education, healthcare, cybersecurity, autonomous vehicles, e-commerce, and customer service, to reveal their real-world impact. Furthermore, we demonstrate some LLM limitations using DeepSeek-R1 and GPT-4o. To the best of our knowledge, this is the first comprehensive study to integrate the challenges and applications of LLMs and Agentic AI within a single forward-looking research landscape that promotes interdisciplinary research and responsible advancement of this emerging field.

In the Netherlands inter-organizational networks have been established to drive long-term healthcare improvement on a regional and national scale. Managing healthcare networks is challenging, especially in sustaining participants' active engagement. This study aims to empirically explore why participants' network commitment declined and how network commitment was managed by network coordinators. We conducted an exploratory qualitative study using purposive sampling we held semi-structured interviews with representatives from 18 Dutch healthcare improvement networks. Respondents mentioned that allocated time and continued participant commitment was required to reach their goals. A key challenge was the decline in participant commitment over time due to misalignment of goals and needs, slow progress, lack of reciprocity, and resistance to change at the organizational level. Approaches to manage this were collaborative reassessment of goals and needs, expectation management, setting preconditions and fostering personal relationships. Our findings highlight the importance of strategic timing of managerial approaches and their interdependencies with social mechanisms. These insights contribute theoretically to Ring and Van de Ven's framework on developmental processes of cooperative interorganizational relationships and offer practical guidance for network managers aiming to maintain long-term commitment.

ContextIn the Netherlands inter-organizational networks have been established to drive long-term healthcare improvement on a regional and national scale. Managing healthcare networks is challenging, especially in sustaining participants’ active engagement. This study aims to empirically explore why participants’ network commitment declined and how network commitment was managed by network coordinators.MethodsWe conducted an exploratory qualitative study using purposive sampling we held semi-structured interviews with representatives from 18 Dutch healthcare improvement networks.ResultsRespondents mentioned that allocated time and continued participant commitment was required to reach their goals. A key challenge was the decline in participant commitment over time due to misalignment of goals and needs, slow progress, lack of reciprocity, and resistance to change at the organizational level. Approaches to manage this were collaborative reassessment of goals and needs, expectation management, setting preconditions and fostering personal relationships.DiscussionOur findings highlight the importance of strategic timing of managerial approaches and their interdependencies with social mechanisms. These insights contribute theoretically to Ring and Van de Ven’s framework on developmental processes of cooperative interorganizational relationships and offer practical guidance for network managers aiming to maintain long-term commitment.

The integration of science, education, and industry has become a prevailing philosophy in higher education. This educational philosophy expands the talent cultivation domain of universities and aligns more closely with the growth patterns of innovative talents, making it an inevitable choice for cultivating high-level innovative talents in China. This study aims to explore the internal logic and mechanism construction path of empowering high-quality innovative talent cultivation through the integration of science, education and industry, in response to the needs of higher education reform and industrial upgrading. Through theoretical analysis and case study methods, it systematically analyzes the prominent issues currently existing in the integration of science, education, and industry, such as insufficient system design, goal misalignment, and limited coverage, revealing its historical inevitability, theoretical rationality, and practical feasibility in three aspects. The study finds that the key to achieving deep integration is to establish a motivating mechanism for multi-party collaboration, an operational mechanism for efficient resource allocation, a scientific evaluation mechanism, and a policy guarantee mechanism. This research provides a systematic solution for the reform of innovative talent cultivation models in universities, supporting the implementation of the national strategy of revitalizing the country through science and education, and offering important guidance for promoting the deep coupling of the education chain, talent chain, and industrial chain.

AimInternship programs are important components of teaching and learning that provide medical students with opportunities for real-life learning. The study aimed at evaluating the internship program for students of xxx using the CIPP model (Context, Input, Process, and Product).Materials and methodsA cross-sectional descriptive analysis was performed on 305 students and 15 faculty members of xxx. Data were collected using a questionnaire based on the CIPP model that developed by researchers. The internship program was evaluated in four areas: input, context, process and product from the perspective of students and faculty members. The scores obtained for each domain were analyzed using SPSS v.21.ResultsContext Evaluation: Significant stakeholder discrepancy (p < 0.001): 78% of students reported inadequate environmental needs assessment vs. 35% of faculty 72% students identified goal misalignment with clinical realities vs. 28% faculty. Input Evaluation: No significant difference (p = 0.32): Comparable ratings for resource adequacy (students: 4.1/6, faculty: 4.3/6) Similar perceptions of curriculum design quality. Process Evaluation: Major implementation gaps (p < 0.001): Supervision quality: Students 2.8/6 vs. Faculty 4.7/6 Feedback mechanisms: 65% student dissatisfaction. Product Evaluation: Strong CIPP domain correlations: Context→Input: r = 0.769 (p < 0.001). Context→Process: r = 0.733 (p < 0.001). Context→Product: r = 0.724 (p < 0.001).ConclusionsThe design and implementation of evaluation programs based on the CIPP model may help improve internship programs and achieve students’ professional competencies. The positive and negative findings in this study should be considered by decision makers and healthcare officials when designing and implementing internship programs. Further longitudinal studies may be required to confirm these findings.Clinical trial numberNot applicable.

ABSTRACT This study examines an urban community gardening initiative with stated ambitions to contribute to both environmental and social sustainability in Stockholm, Sweden, through the lens of postapocalyptic environmentalism. Based on ethnographic fieldwork and qualitative interviews with gardeners as well as local officials, the analysis investigates the links between gardeners’ postapocalyptic motivations, their practical work, and the wider institutional context. Rather than viewing postapocalyptic environmentalism as emerging outside of traditional institutions, the analysis reveals the initiative’s entanglement and co-dependence on institutional structures and policies. The high level of institutionalization of urban community gardening locally requires the initiative to navigate the ‘gray zone’ between autonomy and co-optation and balancing between building ‘small hope’ through gardening work with actively confronting institutional challenges. Vague municipal guidelines, one-size-fits-all regulations and limited resources make urban community gardening in this setting vulnerable to participation fatigue as well as goal misalignment and co-optation.

Abstract Hackathons are intensive design experiences during which teams identify a problem and rapidly develop prototypes of solutions. These events offer a promising venue for studying collaborative design: they are naturalistic, short term, and contained, which mitigates many of the drawbacks of traditional investigations in design research. The objective of this technical brief is to present and describe a transcript dataset of conversations collected from a hackathon team. The dataset includes a transcript of the verbal communication between a four-person hackathon team in the first 2 hours and 9 min of their collaboration. This portion of the design process, totaling 908 segments of speech, details the team’s problem exploration, brainstorming, idea selection, and premature team dissolution. This brief outlines the advantages of having rich transcript data freely available, with a specific focus on new research directions and impact. We include a qualitative analysis of the premature team dissolution, using inductive coding to explore goal misalignment, to provide further context of the collaboration captured in the transcript. We aim for this brief to encourage the use of these data for future investigations of design, teamwork, and hackathon phenomena, as well as act as an exemplar for future publications of open-access datasets.

Modern business environments require organizations to possess advanced tools which Leadership Development Programs (LDPs) develop through their pivotal role in this process. This research examines how Leadership Development Programs (LDPs) support the development of organizational resilience as well as strengthen change management capabilities. An organization needs resilience to respond to disruptions and requires effective change management for achieving smooth transitions during times of essential transformation. Organizational performance metrics have been assessed using a mixed methods model which combines quantitative survey data with qualitative data analysis for evaluating LDP impact. Leadership development programs that follow well-defined structures produce remarkable improvements in leaders' ability to predict future crises as well as their management and adaptation skills. Within organizations LDPs set the stage for adaptable workplaces through which innovation can thrive because team members experience less resistance during transitions. Industrial case studies prove that companies which implement leadership development efforts achieve swift returns through operational efficiency while retaining workforce commitment and maintaining market superiority. Multiple barriers exist in the way of LDP program success including resource limitations and goal misalignment and inconsistent program assessment approaches. To achieve maximal results from LDPs organizations must develop specific leadership development programs that include all stakeholders and connect directly to organizational goals. This research adds value to established literature through its combination of resilience theory with change management theory as precise foundations for leadership development models. This work presents practical guidance to professionals by outlining step-by-step solutions for designing LDPs together with methods to enhance organizational adaptiveness and valid assessment metrics. Organizations must establish leadership development as their strategic base for sustainable growth since changes and uncertainties demand immediate attention

From greenwashing to genuine sustainability: Insights from FinTech and banking executives in emerging market experience.

Real-world case studies offer crucial insights into how corporate–non-profit partnerships generate sustainable social impact, particularly in developing contexts. This article examines five high-profile collaborations—Coca-Cola and WaterAid in Africa, Unilever’s Project Shakti in India, Microsoft and the Akshaya Patra Foundation, Nestlé and the IFRC, and GE with Safe Water Network in Ghana. These examples reveal how shared values, robust communication, and mutual benefits foster enduring impact in areas such as water access, women’s empowerment, nutrition, and disaster relief. The article also highlights lessons from challenges, including misalignment of goals and weak communication, offering best practices such as partner research, clear value propositions, monitoring frameworks, and adaptive management. For practitioners, the findings provide actionable recommendations to structure effective partnerships; for academics, they demonstrate the interplay between cross-sector governance, sustainability, and community-centered approaches.

Background Although medications for opioid use disorder (MOUD) are effective for treating opioid use disorder (OUD), persistent barriers still prevent patients from accessing this life-saving care. Policies to increase MOUD access have produced suboptimal results. This study presents a qualitative system dynamics model that elucidates the complexities of accessing and staying in MOUD treatment. Methods We utilized a community-based system dynamics approach to modeling the MOUD treatment system. We engaged a cohort of system experts/stakeholders, including individuals who had received MOUD, treatment providers, and policymakers, in interviews and group model building to develop and refine a simulation model. We then created a qualitative causal loop diagram based on insights gained while developing the simulation model and a review of interview transcripts. Results The causal loop diagram captures four key factors affecting treatment initiation, retention, and leaving: (1) fraught interactions between patients and healthcare providers; (2) stigma-driven regulation of MOUD creating a culture of fear and defensive medicine; (3) a punitive culture in clinics and opioid treatment programs offering MOUD; and (4) the internalization of the abstinence narrative contributing to premature termination of treatment. Conclusions Our analysis highlights how interdependent and non-linear feedback processes diminish or counteract the effectiveness and sustainability of MOUD policy interventions. Due to system memory and cultural resistance to change, even rolling back reactionary policies may do little to curb established behavioral patterns. In addition, conflicting and competing strategies among various actors within the system contribute to goal misalignment and a lack of standardization of care.

This qualitative study explores supplier collaboration in e-commerce product development, examining motivations, challenges, strategies, outcomes, and ethical considerations. Through in-depth interviews with stakeholders from diverse e-commerce sectors, the study identifies key themes shaping collaboration dynamics. Motivations include innovation, operational efficiency, and strategic partnerships, driving companies to integrate supplier expertise early in product development. Challenges such as global supply chain complexities, goal misalignment, and communication barriers underscore the need for robust governance and cultural sensitivity. Strategies for success include technology adoption, supplier development programs, and collaborative decision-making, enhancing supply chain visibility and mutual benefits. Effective collaboration yields improved product quality, cost efficiencies, and enhanced customer satisfaction, supporting competitive advantage. Ethical sourcing practices and sustainability initiatives are crucial for maintaining trust and regulatory compliance. Cultural and organizational factors, including leadership support and change management, significantly influence collaboration outcomes. The study concludes with implications for theory and practice, emphasizing the role of innovative strategies and continuous improvement in supplier relationships. Future research could explore digital transformation, sustainability trends, and technological impacts on supplier collaboration in e-commerce. Practical applications include optimizing supply chain strategies to navigate complexities and capitalize on market opportunities.