Search Results (108)

Background: Agentic artificial intelligence (AI) systems, characterized by autonomous goal-directed behavior, multi-step reasoning, task decomposition, and tool use, are increasingly proposed for healthcare applications. However, their autonomy raises concerns regarding transparency, accountability, and human oversight. While explainable AI (XAI) has been widely studied in traditional predictive models, less is known about how explainability is implemented within agentic architectures. Objective: To map the emerging literature on explainable agentic AI (XAAI) in healthcare and characterize the types, scope, and forms of explainability used in these systems. Methods: A scoping review was conducted following PRISMA-ScR guidelines. PubMed, Embase, IEEE Xplore, and ACM Digital Library were searched through November 2025. Eligible studies described healthcare-related agentic AI systems incorporating explicit explainability mechanisms. Data were extracted on system architecture, explainability type (intrinsic, post hoc, hybrid), explanation scope (local, global), explanation form, and reported clinical outcomes. Results: Nine studies met the inclusion criteria. All systems demonstrated core agentic features, including autonomy, task decomposition, and tool integration, often within multi-agent frameworks. Explainability was predominantly intrinsic and workflow-native, typically delivered through textual reasoning traces and example-based grounding in retrieved clinical evidence. Feature-based and global explanations were comparatively rare and largely confined to hybrid architectures. Across domains including radiology, neurology, psychiatry, and biomedical research, XAAI systems were reported to improve performance and interpretability relative to baseline models in the included studies. However, these findings were derived from heterogeneous, predominantly experimental or retrospective studies, and structured human-in-the-loop oversight was infrequently described. Conclusions: Current XAAI systems appear to emphasize process transparency and evidence grounding rather than mechanistic model-level attribution. The available evidence remains limited and heterogeneous, and findings should be interpreted as early trends rather than established characteristics. Further progress will require standardized evaluation frameworks, clearer reporting of oversight mechanisms, and validation in real-world clinical settings to support safe and trustworthy integration of agentic AI into healthcare practice. Full article

Traditional intrusion detection systems for IoT networks achieve high classification accuracy but lack interpretability and actionable incident-response capabilities, limiting their operational value in security-critical environments. This paper presents a graph-based multi-agent framework that integrates ensemble machine learning with Large Language Model (LLM)-powered incident report generation via Retrieval-Augmented Generation (RAG). The system employs a three-phase architecture: (1) a lightweight Random Forest binary pre-detection, achieving 99.49% accuracy with a 6 MB model size for edge deployment; (2) ensemble classification combining Multi-Layer Perceptron, Random Forest, and XGBoost with soft voting and SHAP-based feature attribution for explainability; and (3) a ReAct-based summary agent that synthesizes classification results with external threat intelligence from Web search and scholarly databases to generate evidence-grounded incident reports. To address the challenge of evaluating non-deterministic LLM outputs, we introduce custom RAG evaluation metrics—faithfulness and groundedness implemented via the LLM-as-Judge framework. Experimental validation on the ACI IoT Network Dataset 2023 demonstrates ensemble accuracy exceeding 99.8% across 11 attack classes; perfect groundedness scores (1.0), indicating all generated claims derive from the retrieved context; and moderate faithfulness (0.64), reflecting appropriate analytical synthesis. The ensemble approach mitigates individual model weaknesses, improving the UDP Flood F1 score from 48% (MLP alone) to 95% through soft voting. This work bridges the gap between high-accuracy detection and trustworthy, actionable security analysis for automated incident-response systems. Full article

With the widespread adoption of anonymity networks such as Tor, I2P, and JonDonym, reliably classifying darknet traffic remains challenging due to feature redundancy and severe class imbalance in encrypted flows. Existing approaches often rely on static feature-selection strategies and generic oversampling methods, which limit robustness and may distort traffic semantics. This study proposes an adaptive classification framework integrating Adaptive Weighted Feature Aggregation (AWFA) for reliability-aware feature selection and Traffic-Aware SMOTE (TA-SMOTE) for semantically constrained perturbations of packet-size and timing features while preserving flow-level structure. The framework is evaluated on a two-layer hierarchy comprising browser-level (L1) and application-level (L2) classification. At the L2, the proposed AWFA and TA-SMOTE pipeline attains a macro-F1 score of 73.81%, significantly exceeding PCA-based reduction and traditional RF-based selection with SMOTE. At the browser level (L1), macro-F1 rises from 91.58% to 96.09% while reducing the feature space from 84 to 40 attributes, highlighting both performance improvements and structural efficiency gains. Additional semantic validation confirms that the balancing process preserves the statistical and structural characteristics of genuine darknet traffic. These results indicate that reliability-aware feature aggregation and traffic-aware balancing provide a practical, trustworthy approach to modern darknet traffic classification. Full article

Reinforcement learning (RL) agents are increasingly deployed for voltage control in power distribution networks. However, their opaque decision-making creates a significant trust barrier, limiting their adoption in safety-sensitive operational settings. This paper presents XRL-LLM, a novel framework that generates natural language explanations for RL control decisions by combining game-theoretic feature attribution (KernelSHAP) with large language model (LLM) reasoning grounded in power systems domain knowledge. We deployed a Proximal Policy Optimization (PPO) agent on an IEEE 33-bus network to coordinate capacitor banks and on-load tap changers, successfully reducing voltage violations by 90.5% across diverse loading conditions. To make these decisions interpretable, KernelSHAP identifies the most influential state features. These features are then processed by a domain-context-engineered LLM prompt that explicitly encodes network topology, device specifications, and ANSI C84.1 voltage limits.Evaluated via G-Eval across 30 scenarios, XRL-LLM achieves an explanation quality score of 4.13/5. This represents a 33.7% improvement over template-based generation and a 67.9% improvement over raw SHAP outputs, delivering statistically significant gains in accuracy, actionability, and completeness ($p<0.001$, Cohen’s d values up to 4.07). Additionally, a physics-grounded counterfactual verification procedure, which perturbs the underlying power flow model, confirms a causal faithfulness of 0.81 under critical loading. Finally, five ablation studies yield three broader insights. First, structured domain context engineering produces synergistic quality gains that exceed any single knowledge component, demonstrating that prompt composition matters more than the choice of foundational model. Second, even an open source 8B-parameter model outperforms templates given the same prompt, confirming the framework’s backbone-agnostic value. Most importantly, counterfactual faithfulness increases alongside load severity, indicating that post hoc attributions are most reliable in the high-stakes regimes where trustworthy explanations matter most. Full article

With the increasing deployment of autonomous driving systems, the opaque nature of deep reinforcement learning (DRL) decision models hinders understanding and validation of driving decisions. To address this challenge, we propose a Hybrid Attribution-based Interpretable Deep Reinforcement Learning framework (HA-IDRL) for autonomous driving behavior decision-making. The framework introduces a Hybrid Gradient–LRP (HGL) attribution mechanism that integrates gradient-based attribution and Layer-wise Relevance Propagation (LRP) to capture complementary sensitivity and contribution information, producing more consistent and comprehensive post hoc explanations. In addition to post hoc interpretability, we enhance structural interpretability by replacing the conventional multilayer perceptron (MLP) in the Dueling Deep Q-Network (Dueling DQN) architecture with Kolmogorov–Arnold Networks (KAN). By representing nonlinear interactions through learnable univariate functions and explicit summation structures, KAN provides inherently interpretable functional decompositions. The proposed framework is evaluated on a highway lane-changing task using the highway-env simulator. Experimental results show that HA-IDRL achieves decision-making performance comparable to representative DRL baselines, including Dueling DQN and Soft Actor-Critic (SAC), while providing explanations that are more stable and better aligned with human driving semantics. Moreover, the proposed method produces explanations with low computational overhead, enabling efficient and real-time interpretability in practical autonomous driving applications. Overall, HA-IDRL advances trustworthy autonomous driving by enabling high-performance decision-making and rigorous, multi-level interpretability, thereby improving the transparency and operational reliability of DRL-based driving policies. Full article

We propose Trust Triangle, a Bridging Methodology that establishes evidential reliability through multi-attribution consensus, ensures external validity via statistical hypothesis testing, and enables controlled generation with RAG-anchored LLMs to transform black-box predictions into trustworthy, auditable explanations. This framework is instantiated for credit card fraud detection by integrating multi-method feature attributions with rigorous statistical validation. The resulting reliability-validity-verified insights are synthesized with high-relevance domain knowledge (relevance score > 0.7) retrieved from a real-world corpus via Retrieval-Augmented Generation (RAG). A structured Chain-of-Thought (CoT) prompt then guides an LLM to produce coherent, audit-ready case reports. Our contributions are threefold: (1) a verifiable framework for quantifying attribution reliability and validity, (2) a demonstrated end-to-end pipeline from robust prediction to semantically grounded explanation, and (3) a generalizable paradigm for Trustworthy ML in high-stakes domains. Experiments on a highly imbalanced dataset (fraud rate: 8.74%) demonstrate robust performance (PR-AUC = 0.7867), successfully identify statistically significant predictive features, and generate audit-ready reports, thereby advancing a rigorous, evidence-based pathway from model output to decision-ready support. Full article

Conventional topical therapy for corneal and anterior segment diseases is limited by rapid tear clearance and multilayer corneal barriers, resulting in low bioavailability and the need for frequent dosing. Artificial intelligence (AI) is emerging as a complementary approach that learns quantitative relationships between molecular structure, formulation variables, and ocular performance. In corneal drug delivery, machine learning models have been used to optimize multicomponent formulations and processing conditions; predict key quality attributes such as particle size, zeta potential, encapsulation efficiency and release kinetics; and estimate corneal permeability, retention and ocular irritation risk, thereby reducing experimental burden and guiding safer design. AI can also be coupled with mechanistic ocular pharmacokinetic/pharmacodynamic models to translate formulation attributes into predicted tissue exposure. Finally, inverse design approaches enable the discovery of new carriers and devices, illustrated by machine learning-guided peptide carriers and smart contact lens platforms that combine sensing with on-demand drug release. Despite these advances, current datasets remain small and heterogeneous, external validation and benchmarking against conventional workflows are limited, and uncertainty quantification and interpretability must be addressed to enable clinical translation. This review summarizes corneal barriers and delivery platforms, critically evaluates where AI provides measurable value across design, characterization and performance and highlights data and validation priorities needed for trustworthy AI-enabled corneal therapeutics. Full article

Hospitals are energy- and waste-intensive systems. Inpatient buildings dominate the sector’s electricity and gas consumption, and healthcare waste streams—especially device-associated disposables—increase environmental burdens. AI-enabled predictive maintenance (PdM) offers a dual lever: (1) reducing energy use by keeping assets operating at efficient points, and (2) preventing avoidable waste by extending component life, reducing emergency spares, and avoiding device-induced clinical workflow disruptions. In this study, an end-to-end architecture is developed by integrating multi-modal sensing (electrical, thermal, acoustic, vibration), computerized maintenance management systems (CMMS), risk-based maintenance under International Electrotechnical Commission (IEC)/International Organization for Standardization standards (ISO 60601, 62353/62304, 81001-5-1), and learning pipelines (self-supervised anomaly detection, remaining useful life estimators, and carbon-aware work order scheduling). Using representative hospital archetypes and equipment classes (imaging, patient monitoring, laboratory analyzers, sterilizers, and pumps), energy, downtime, and waste avoidance are simulated under baseline preventive maintenance (PM) versus PdM with alternate equipment management. Results showed that 10–22% site electricity reduction was achieved, attributable to equipment efficiency and optimized duty-cycling, 18–35% fewer unplanned failures, and a 12–28% reduction in associated consumable waste and emergency part scrappage across scenarios, while maintaining compliance with Joint Commission/Centers for Medicare & Medicaid Services and IEC safety testing intervals. We discuss cybersecurity (IEC 81001-5-1) and the trustworthiness of AI, present a governance model linking CMMS events to carbon telemetry, and provide an implementation roadmap. Full article

The rapid diffusion of “(GenAI)” Generative Artificial Intelligence systems has reshaped everyday activities, yet their adoption remains uneven and cognitively demanding for many users. Existing research has largely relied on conventional technology acceptance models, providing limited insight into cognitive burden and GenAI-specific system characteristics. To address this gap, this study develops an integrated framework combining the Technology Acceptance Model, Cognitive Load Theory, and the DeLone and McLean Information Systems Success Model to explain GenAI adoption among ordinary users. Survey data from 1001 active GenAI users were analyzed using partial least squares structural equation modeling (PLS-SEM). The results indicate that all core technology acceptance relationships are statistically significant (p < 0.001), while mental load negatively affects perceived usefulness and user attitudes. Moreover, GenAI system attributes—output quality, transparency, friction reduction, and system integration—significantly moderate key adoption pathways and strengthen the translation of behavioral intention into actual use. Predictive assessment indicates that the proposed model outperforms the baseline technology acceptance model, with stronger explanatory power and superior out-of-sample predictive performance (Q²predict > 0.35). The findings offer actionable insights for designing cognitively efficient, trustworthy, and sustainable GenAI systems. Full article

Due to the rapid spread of disinformation, it is becoming increasingly difficult for the public to understand current events and how discussions and decisions are made in democratic societies. We propose a KG-grounded narrative extraction pipeline to compare disinformation and trustworthy news. English articles (2015–2023), included in EUvsDisinfo cases and matched mainstream coverage, were converted to AMR-based RDF graphs, and LLM ensembles were used to extract characters, events, causal links and framing edges grounded in these graphs. We studied two ensemble policies: a recall-oriented union that retained all model outputs and a precision-oriented consensus that kept only agreed elements, plus an LLM critic that flagged missing links, contradictions and framing inconsistencies. On an expert-annotated subset of 60 articles, the extractor ensemble attained very high precision for characters (0.99) and events (0.97) and solid performance for causal links (0.77) and framing edges (0.84), with similar scores for both classes. Our critic ensemble reached 0.74 precision. Structurally, union and consensus operated over the same grounded nodes but differed significantly in relational density, thus achieving rich vs. skeletal narrative graphs. Linking our narratives to GDELT showed that $\approx 97\%$ of extracted actors and events appeared in global news for both classes, while directional actor pairs from causal links were less often supported for disinformation (0.45) than trustworthy news (0.60). Overall, disinformation and trustworthy articles shared event backbones but diverged in the density and (to a lesser extent) directionality of causal attributions and framing relations. Full article

Professionalism is central to veterinary practice, shaping not only the quality of care provided to animals but also the wellbeing of practitioners, the satisfaction of clients, and the sustainability of the profession. Prior research has catalogued various attributes of professionalism that are important for career success, but few studies have integrated these multiple perspectives into a cohesive framework. This study synthesizes insights from three key veterinary stakeholder groups—students, clinical practitioners, and clients—using a multi-methods approach including surveys, focus groups, critical incident interviews, and client complaint analyses. Across the datasets, ranking of Likert-scale responses and thematic analysis revealed four recurring themes that were identified as essential for career success: ‘Effective communication’; ‘Accountability, integrity, trustworthiness, and honesty’; ‘Personal wellbeing’; and ‘Quality of service’. These themes were organized into a unifying theoretical model of veterinary professionalism, conceptualized as a ‘navigational compass’, comprising three domains of care: patient-centered care, relationship-centered care, and self-care. By conceptualizing professionalism in terms of a compass, the model illustrates how veterinarians can draw on key professionalism attributes, coupled with consideration of the three domains of veterinary care, to navigate the challenges of practice and sustain long-term career success. The compass provides a reflective framework to guide veterinarians and educators, to support the integration of professionalism into curricula and to guide careers toward excellence in care and lasting personal fulfilment. Full article

Artificial intelligence (AI) powers breakthroughs in language processing, computer vision, and scientific discovery; yet, the increasing complexity of frontier models makes their reasoning opaque. This opacity undermines public trust, complicates deployment in safety-critical settings, and frustrates compliance with emerging regulations. In response to initiatives such as the White House AI Action Plan, we synthesize the scientific foundations and policy landscape for interpretability, control, and robustness. We clarify key concepts and survey intrinsically interpretable and post-hoc explanation techniques, discuss human-centered evaluation and governance, and analyze how adversarial threats and distributional shifts motivate robustness research. An empirical case study compares logistic regression, random forests, and gradient boosting on a synthetic dataset with a binary-sensitive attribute using accuracy, $F_1$ score, and group-fairness metrics, and illustrates trade-offs between performance and fairness. We integrate ethical and policy perspectives, including recommendations from America’s AI Action Plan and recent civil rights frameworks, and conclude with guidance for researchers, practitioners, and policymakers on advancing trustworthy AI. Full article

Accurate and trustworthy forecasting of wind and photovoltaic power generation is essential for the reliable operation and planning of modern power systems. Although recent machine-learning-based forecasting solutions increasingly incorporate elements of trustworthy artificial intelligence, such as explainability, uncertainty quantification, robustness, drift monitoring, and machine learning operations, these components are typically introduced in a fragmented manner and remain weakly integrated at the architectural level, which limits their applicability in real operational environments. This paper presents a systematic review of 59 peer-reviewed journal articles published between 2019 and 2025, conducted in accordance with the PRISMA 2020 guidelines. The review includes studies focused on wind and photovoltaic power forecasting that report system architectures, frameworks, or end-to-end pipelines incorporating at least one trust-related attribute. The literature search was performed using Scopus, IEEE Xplore, MDPI, and ScienceDirect. Using a narrative and architectural synthesis, the review identifies six structural gaps hindering industrial deployment: the absence of semantic data models, shallow model-centric explainability, drift monitoring without governance mechanisms, lack of automated model lifecycle management, insufficient robustness to real-world data defects, and the absence of integrated end-to-end architectures. The evidence base is limited by the heterogeneity of architectural descriptions and the predominantly qualitative nature of reported implementations. Based on these findings, a high-level reference architecture for a trustworthy AI-based forecasting system is proposed. The architecture formalizes trustworthiness as a system-level property and integrates semantic, technological, and functional trust layers within a unified data and model lifecycle, supporting reproducible, interpretable, and operationally reliable forecasting for both wind and photovoltaic power plants. Full article

The increasing integration of artificial intelligence (AI) in decision-making processes has amplified discussions surrounding algorithmic authority—the perceived epistemic legitimacy of AI systems over human judgment. This study investigates how individuals attribute epistemic authority to AI, focusing on psychological, contextual, and sociotechnical factors. Existing research highlights the importance of trust in automation, perceived performance, and moral frameworks in shaping such attributions. Unlike prior conceptual or philosophical accounts of algorithmic authority, our study adopts a relational and empirically grounded perspective by operationalizing algority through psychometric measures and contextual assessments. To address knowledge gaps in the micro-level dynamics of this phenomenon, we conducted an empirical study using psychometric tools and scenario-based assessments. Here, we report key findings from a survey of 610 participants, revealing significant correlations between trust in automation (TiA), perceptions of automated performance (PAS), and the propensity to defer to AI, particularly in high-stakes scenarios like criminal justice and job-matching. Trust in automation emerged as a primary factor, while moral attitudes moderated deference in ethically sensitive contexts. Our findings highlight the practical relevance of transparency and explainability for supporting critical engagement with AI outputs and for informing the design of contextually appropriate decision support. This study contributes to understanding algorithmic authority as a multidimensional construct, offering empirically grounded insights for designing AI systems that are trustworthy and context-sensitive. Full article

As artificial intelligence (AI) is increasingly used in classrooms, tutoring systems, and learning platforms, it is essential that these tools are not only powerful, but also easy to understand, fair, and supportive of real learning. Many current AI systems can generate fluent responses or accurate predictions, yet they often fail to clearly explain their decisions, reflect students’ cultural contexts, or give learners and educators meaningful control. This gap can reduce trust and limit the educational value of AI-supported learning. This paper introduces the PEARL framework, a human-centered approach for designing and evaluating explainable AI in education. PEARL is built around five core principles: Pedagogical Personalization (adapting support to learners’ levels and curriculum goals), Explainability and Engagement (providing clear, motivating explanations in everyday language), Attribution and Accountability (making AI decisions traceable and justifiable), Representation and Reflection (supporting fairness, diversity, and learner self-reflection), and Localized Learner Agency (giving learners control over how AI explains and supports them). Unlike many existing explainability approaches that focus mainly on technical performance, PEARL emphasizes how students, teachers, and administrators experience and make sense of AI decisions. The framework is demonstrated through simulated examples using an AI-based tutoring system, showing how PEARL can improve feedback clarity, support different stakeholder needs, reduce bias, and promote culturally relevant learning. The paper also introduces the PEARL Composite Score, a practical evaluation tool that helps assess how well educational AI systems align with ethical, pedagogical, and human-centered principles. This study includes a small exploratory mixed-methods user study (N = 17) evaluating example AI tutor interactions; no live classroom deployment was conducted. Together, these contributions offer a practical roadmap for building educational AI systems that are not only effective, but also trustworthy, inclusive, and genuinely supportive of human learning. Full article