Search

In recent years, considerable attention has been directed toward clarifying complex, opaque systems, commonly referred to as black-box models, including deep neural networks and random forests. So-called model-agnostic techniques generally approximate the outputs of these non-transparent systems using an interpretable architecture, ignoring the internal mechanisms of the primary black-box system itself. Consequently, it remains an important inquiry whether the straightforward training of transparent, white-box architectures ought to be favored over the post-hoc approximation of these opaque models. This study presents the findings of an empirical investigation comparing post-hoc explanatory methods against inherently interpretable architectures across multiple datasets, focusing on both rule-based and feature-based frameworks. The empirical outcomes indicate that directly trained transparent models frequently approximate black-box models with an accuracy that matches or exceeds that of post-hoc surrogate models, despite lacking direct access to the underlying black-box structure.

Identifying vehicles within parking zones yields valuable insights into the spatial and temporal distribution of parking spot utilization. Because parking monitoring is conventionally conducted through manual observation, its temporal granularity remains restricted by substantial personnel expenditures. This study leverages simulated datasets alongside transfer learning to construct a dependable, real-world model tasked with detecting and categorizing vehicles via single-beam LiDAR within a roadside parking context. This work introduces a synthetically augmented transfer learning methodology designed for LiDAR-centric vehicle detection, alongside the deployment framework for synthetic LiDAR datasets. A synthetically augmented transfer learning technique was implemented to expand the constrained real-world dataset and facilitate the formulation of data management workflows. Furthermore, incorporating this synthetically augmented transfer learning methodology enhances both the resilience and the overall precision of the network. Empirical tests demonstrate that this approach enables the rapid deployment of models tailored for vehicle detection using a LiDAR sensor.

The integration of deep learning (DL) algorithms offers significant potential to enhance healthcare applications. Specifically, DL has advanced medical imaging diagnostics, therapeutic interventions, and disease management. However, deploying deep learning algorithms on sensitive medical imagery introduces substantial challenges regarding data security and patient privacy. Optimizing medical imaging performance while simultaneously ensuring patient anonymity remains a complex task. Consequently, privacy-preserving techniques designed for deep learning model training and inference are increasingly gaining traction. Currently, these medical image sequences are interpreted using state-of-the-art computer-aided detection/diagnosis (CAD) methodologies. Yet, frameworks that require uploading medical photographs to external servers inherently create privacy vulnerabilities.To address this, this article introduces a convolutional Bi-LSTM network designed to evaluate fully homomorphic-encrypted (HE) time-series medical images. Within this framework, convolutional blocks extract selective spatial features from the secure image sequences, while analytical sequence layers based on Bi-LSTM encode the temporal data. To optimize efficiency and minimize misdiagnoses, a weighted unit and sequence voting layer incorporates spatial data using varying weights. The effectiveness of this framework is demonstrated using two rigorous public benchmarks: the CheXpert and BreaKHis datasets. The proposed technique surpasses several competing methods, achieving an accuracy exceeding 0.99 across both datasets. These findings indicate that the introduced methodology can successfully extract visual representations and sequential dynamics from encrypted medical image series, thereby safeguarding patient privacy while maintaining high performance in medical image analysis.

The climate system is heavily influenced by aerosols through direct and indirect pathways, notably the absorption and scattering of radiant energy. Beyond climate impacts, these particles compromise atmospheric visibility and jeopardize public health. Specifically, human populations face exposure to fine particulate matter PM2.5, a major driver of respiratory and cardiovascular pathologies. Long-term modifications in PM levels are generally governed by emission shifts and meteorological patterns, whereas short-term fluctuations are predominantly dictated by immediate weather dynamics. Surface variations in PM2.5 levels are further driven by environmental variables such as temperature, wind velocity, relative humidity, and vegetative coverage.Air quality forecasting has advanced significantly through the application of machine learning. This research explores PM2.5 estimation by utilizing these environmental and meteorological metrics as spatiotemporal inputs. This study evaluates seven years (2014–2021) of in situ PM2.5 measurements gathered from five distinct monitoring stations across Singapore, establishing a baseline dataset for spatial modeling.The primary contribution of this paper is a novel predictive architecture tailored for temporal analysis, utilizing Random Forest (RF), Gradient Boosting (GB) regression, and the Tree-based Pipeline Optimization Tool (TP) Auto ML framework, which optimizes via a meta-heuristic genetic algorithm. The TP model demonstrated strong predictive capabilities, achieving a peak Global Performance Index (GPI) of 7.4 in August 2016 and a minimum value of −0.6 in June 2019. These metrics highlight the robust nature of the TP framework; its negative values remained superior to competing models, indicating less pessimistic errors. Finally, the underlying logic of these machine learning systems is clarified using eXplainable Artificial Intelligence (XAI) methods, which validate the integrity of feature importance and offer deeper insights into the cyclical variations of PM2.5 concentrations.

The integration of digital architectures powered by artificial intelligence (AI) into daily routines is expanding at an unprecedented rate. Within this paradigm, the twin dilemmas of accountability and openness endure as critical issues that have yet to be satisfactorily resolved. This study intends to delineate the specific domains of opacity within digital ecosystems and formulate actionable frameworks for mitigating these visibility challenges. Initially, by synthesizing existing scholarship, interrogating the dynamics of platformization, establishing a cross-disciplinary conceptualization of AI, and applying ethical AI frameworks, we map the dimensions of opacity inherent in digital environments. Subsequently, having mapped these obscure domains, we move beyond a standard literature synthesis to offer an independent perspective on mitigating these systemic vulnerabilities. Our evaluation of the literature indicates the existence of three primary spheres of opacity within digital networks: an inadequate understanding of platform content originators; a deficiency of visibility regarding the background labor force, their explicit operational outputs, and their occupational environments; and a lack of openness concerning the engineering and oversight of algorithmic models. Given the abundance of abstract, high-level ethical tenets in current scholarship that resist pragmatic implementation, this investigation represents an effort to reconcile theoretical concepts with operational execution.

We introduce and execute a transparent machine learning classification model designed for Explainable AI (XAI) that leverages expressive Boolean formulas. Potential domains of use include credit risk evaluation and medical diagnostics. The system utilizes a Boolean formula to establish a rule with adjustable complexity (or interpretability) for categorizing input data. This formula accommodates any operator applicable to single or multiple Boolean variables, offering superior expressivity relative to more rigid rule- and tree-based methodologies. Native local optimization methods are used to train the classifier, allowing for an efficient search of the feasible formula space. Fast Integer Linear Programming (ILP) or Quadratic Unconstrained Binary Optimization (QUBO) solvers—potentially enhanced by custom hardware or quantum systems—can identify shallow rules. By executing non-local moves that optimize across subtrees of the complete Boolean formula, we merge the expressivity and efficiency of the native local optimizer with the rapid processing capabilities of these devices. We present comprehensive numerical benchmarking data against multiple baseline models using recognized public datasets. The findings indicate that the native local rule classifier performs competitively with alternative models. Incorporating non-local moves yields comparable outcomes with a reduced number of iterations. Consequently, deploying specialized or quantum hardware could provide substantial acceleration by rapidly generating non-local moves.

Verbal autopsies (VA) serve as a frequent mechanism within Low- and Medium-Income Countries (LMIC) to ascertain the cause of death (CoD) for individuals who pass away outside of clinical environments, relative to the established international gold standard of physician medical certification. Field interviewers gather detailed testimonies from the families of the deceased concerning the events and circumstances preceding the fatality. This gathered information is documented in text format, referred to as VA narratives. These textual records contain granular details suitable for identifying the CoD. Nonetheless, this methodology continues to be a manual operation characterized by high costs, inconsistency, prolonged processing times, and subjectivity (making it susceptible to errors), among other limitations. Consequently, these disadvantages undermine the VA reporting infrastructure, which is otherwise crucial for defining public health agendas and populating civil registration frameworks.To address this operational gap, the present study introduces interpretable deep learning (DL) methodologies designed to analyze VA narratives and deliver CoD predictions in an efficient, transparent, economical, and precise manner. The verification of our DL models is conducted via optimization and performance accuracy machine learning (ML) curves mapped against the volume of training samples. Our models achieved training set accuracies of (LSTM = 76.11%, CNN = 76.35%, and SEDL = 82.1%), validation accuracies of (LSTM = 67.05%, CNN = 66.16%, and SEDL = 82%), and test set accuracies of (LSTM = 67%, CNN = 66.2%, and SEDL = 82%). Additionally, Local Interpretable Model-agnostic Explanations (LIME) are utilized to enhance the interpretability of the outcomes, thereby fostering confidence in automated healthcare solutions.This work offers robust deep learning strategies for CoD determination via VAs, where the stacked ensemble deep learning (SEDL) model achieved optimal performance, outperforming both Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN). Our empirical evidence suggests that ensemble DL paradigms can be successfully integrated into the CoD classification workflow to assist medical experts in reaching a diagnosis. Ultimately, this integration aims to compress the diagnostic turnaround time for physicians reviewing the narratives, while simultaneously reducing operational expenses and mitigating human error. The scope of this study was constrained by the volume of training data required for model optimization and the substantial lexical diversity present within the unstructured text.

Although Reinforcement Learning (RL) exhibits significant potential for refining intricate control and decision-making systems, Deep Reinforcement Learning (DRL) suffers from a distinct lack of explainability. This limitation restricts its deployment within highly regulated domains, including manufacturing, finance, and healthcare. The opacity of DRL decision-making introduces structural challenges that impede operational efficiency and resource management—an issue that intensifies alongside ongoing algorithmic advancements. While current trends favor transitioning from traditional Experience Replay configurations to A3C architectures, the latter imposes substantially greater resource demands. Furthermore, despite continuous efforts to refine selection mechanisms within Experience Replay, repository capacities generally remain elevated. In this study, we evaluate a Deep Convolutional Q-learning agent across 20 Atari environments while deliberately downsizing the Experience Replay capacity from 1×106 to 5×102. Our findings indicate that lowering the capacity from 1×104 to 5×103 does not meaningfully degrade performance rewards, thereby establishing a viable strategy for resource-efficient DRL frameworks. To clarify agent choices and correlate them with inherent game mechanics, we introduce a novel methodology: visualizing the Experience Replay mechanism utilizing the Deep SHAP Explainer. This framework enhances domain comprehension and delivers transparent, interpretable accounts, though capacity constraints must be managed carefully to mitigate overfitting. Ultimately, this research validates the practicality of minimizing Experience Replay storage and supports the integration of transparent explanations via the Deep SHAP Explainer to advance resource conservation.

Machine learning (ML) techniques have been applied in diverse manners to combat the COVID-19 pandemic. Predictive ML models have been constructed, for example, for both diagnostic and prognostic objectives, utilizing varying data modalities such as textual, visual, and structured formats. Because of the distinct characteristics of this disease and its temporal transformation, a comprehensive understanding of the central elements driving the clinical trajectory of COVID-19 in individual patients remains incomplete. Throughout all phases of this study, a medical specialist was actively engaged via a human-in-the-loop framework. This collaborative element represents a vital yet frequently overlooked component of the ML and knowledge extraction (KE) workflow. Our findings indicate that explainable artificial intelligence (XAI) can offer substantial reinforcement to this dimension of ML and KE. This investigation centered on employing ML for knowledge extraction across two distinct operational scenarios. In the initial scenario, we sought to determine if integrating data regarding the dominant COVID-19 variant affects the predictive performance of the ML models. In the subsequent scenario, we concentrated on prognostic classification models designed to predict a patient's requirement for intensive care unit admission, integrating various XAI methodologies. We implemented nine distinct ML techniques, specifically XGBoost, CatBoost, LightGBM, logistic regression, Naive Bayes, random forest, SGD, SVM-linear, and SVM-RBF. The predictive quality of these models was evaluated using precision, accuracy, and AUC metrics. Following this evaluation, knowledge extraction was performed on the highest-performing models through two discrete pathways: (a) automated feature extraction using forward stepwise selection (FSS); and (b) the identification of attributes and their interactions using model explainability tools. Both sets of features were evaluated against a baseline configuration of attributes selected a priori by medical professionals using domain expertise. Our empirical tests demonstrated that incorporating data on the prevailing COVID-19 variant did not alter the performance of the generated ML models. Furthermore, the results indicated that human medical experts displayed significantly higher precision in identifying meaningful attributes compared to the automated FSS approach. Explainability tools identified nearly identical sets of attributes as the medical expert, while additionally uncovering noteworthy attribute interactions that the specialist subsequently evaluated from a clinical perspective. The outcomes of this investigation and their broader implications are examined.

Machine learning (ML) has evolved into an indispensable technology within the defense sector, driving the creation of sophisticated mechanisms for identifying threats, supporting decision-making processes, and managing autonomous operations. Nevertheless, the expanding integration of ML into defense frameworks has generated significant ethical anxieties concerning accountability, transparency, and inherent bias. This paper offers an extensive evaluation of ML’s influence on the defense industry, detailing both the advantages and risks of implementing these methodologies across diverse domains, including surveillance operations, target pinpointing, and autonomous weapon platforms. Furthermore, we examine the ethical repercussions of deploying ML in military contexts, emphasizing problems tied to privacy preservation, responsibility attribution, and algorithmic bias. Ultimately, we outline strategic recommendations to alleviate these ethical complications, advocating for heightened operational transparency, clearer accountability structures, and the active engagement of diverse stakeholders in the architecture and deployment of ML systems within defense frameworks.

The expansion of machine-learning frameworks into sectors with significant end-user consequences has heightened concerns regarding the equity of trained models, particularly the necessity of preventing biases that disadvantage distinct demographic sub-groups. The majority of current bias-mitigation techniques adjust the relative weight of data observations during the pre-processing phase. Because fairness depends heavily on context, we support an interactive machine-learning strategy that empowers human supervisors to offer iterative guidance for model refinement. To this end, we adapt the explanatory interactive machine-learning framework Caipi for equitable machine learning. This modified approach, FairCaipi, integrates human-in-the-loop interventions regarding both model predictions and underlying explanations to enhance model equity. Empirical outcomes indicate that FairCaipi outpaces a state-of-the-art pre-processing bias mitigation strategy in both the equity metrics and the predictive accuracy of the resulting machine-learning framework. We show that FairCaipi serves to expose and minimize algorithmic bias in machine-learning architectures while simultaneously identifying underlying human bias.

This study develops an expanded framework for extracting classification rules from neural network architectures applied to tabular data. We introduce alternative designs for learning Boolean logic operators by configuring neurons with continuous weights, including the integration of trainable NAND gates. This unified formulation ensures that distinct logical operations are trained under a singular mathematical approach, allowing developers to directly deploy established sparsity-inducing optimization routines—such as reweighted L1 regularization—to simplify the resulting network into concise rule sets. Furthermore, by utilizing De Morgan’s law, we construct an alternative network topology based on these trainable NAND components to create a NAND-NAND structure rather than the traditional AND-OR setup, both of which easily convert into discrete decision rules. Empirical evaluations show that these revised methodologies produce highly competitive decision rule models that match state-of-the-art predictive accuracy on tabular datasets.

Machine learning-driven decision support tools ought to assist users in detecting prospective advantages and risks. Widespread model-independent explanation frameworks can pinpoint elements underpinning specific forecasts—addressing queries like "What factors affect sales?" or "Why did sales decline?"—yet fail to indicate the actionable steps an individual can or should take to achieve a more favorable result. Although counterfactual explanation methods target actionable interventions and occasionally account for feasibility, they generally overlook viability for real-time systems like interactive question-answering platforms. In this study, we bridge this divide by proposing a unique model-agnostic technique that details concrete, viable adjustments capable of altering the predictions of a sophisticated Black Box AI framework for a specific instance; furthermore, we evaluate its practical utility by evaluating its computational speed in real time alongside its capacity to discover executable modifications. The developed approach leverages the target instance to produce high-precision explanations, subsequently employing a secondary mechanism to identify achievable minimally-contrastive counterfactual explanations (AMCC) by restricting the exploration space exclusively to modifications that adhere to domain-specific restrictions. Utilizing a well-established dataset, we assessed the classification procedure to determine both the success rate and the computational duration necessary to pinpoint valid counterfactual alternatives. Within the context of a classifier maintaining 90% accuracy, our proposed algorithm successfully extracted AMCC explanations in 47% of instances (38 of 81), requiring an average generation time of 80 ms. These outcomes confirm the efficiency of the algorithm in rapidly generating AMCC explanations, establishing its appropriateness for real-time deployments. The AMCC approach improves the interpretability of obscure Black Box AI models, thereby assisting stakeholders in analyzing corrective techniques or evaluating prospective consequences.

With the introduction of the Algorithmic Transparency Standard, UK authorities have renewed their emphasis on increasing operational openness for automated tools within public administration. This framework demonstrates that the UK is progressing its artificial intelligence regulatory goals while ensuring these benchmarks are grounded in the practical, real-world experiences of operators to achieve uniform, widespread integration. By utilizing a dual-level reporting structure, the Standard encourages inclusive disclosure tailored to different groups, thereby strengthening confidence among varied institutional and public actors. Furthermore, deploying these protocols across public agencies will likely serve as a blueprint for broader corporate benchmarks, guiding compliance models within commercial markets. This paper delivers a structured analysis and critique of the policy document.

The pressing issue of adversarial vulnerability in classifiers, particularly those implemented via deep neural networks, is examined. This problem is investigated through a generalization applicable to any classification models generated by machine learning techniques. The insufficiency of conventionally accepted metrics for evaluating classifier quality is highlighted, including metrics deployed to validate the efficacy of defense mechanisms against adversarial manipulation. The underlying causes behind the generation of adversarial examples and other machine learning-based classification errors are explored. A methodology for simulating adversarial attacks is introduced, demonstrating the core phenomena encountered during such exploits. The necessity of establishing quality metrics specifically tailored to quantify potential classifier vulnerability to adversarial threats is emphasized. To evaluate robustness against adversarial interference, the application of the multidimensional EDCAP metric (Excess, Deficit, Coating, Approx, Pref) is proposed. Furthermore, a technique is presented for constructing a novel multiclass classifier, termed EnAE (Ensemble of Auto-Encoders), which leverages an ensemble of single-class classifiers governed by the EDCAP evaluation criteria. The EnAE classification framework employs a hard voting mechanism and is capable of identifying anomalous input data. The introduced metric, construction framework, and classifier are validated across multiple datasets characterized by feature spaces of moderate dimensionality.

Conventional authentication mechanisms, such as passwords, are becoming less effective due to technological progress and evolving attack methodologies. To address this vulnerability, continuous authentication has emerged as a viable approach, wherein a user's identity is perpetually monitored after initial login to ensure an unauthorized individual has not hijacked the session. Mouse dynamics, which quantifies the behavioral patterns of a user's cursor movements, represents a highly promising biometric indicator for these ongoing verification models. This paper extends our prior research by benchmarking a 40-user dataset across three machine learning and three deep learning models. Our investigation covers two distinct evaluation frameworks: binary classification for individual authentication—where a 1-dimensional convolutional neural network (1D-CNN) yielded the highest performance, achieving a peak average test accuracy of 85.73% for the top-10 users—and multi-class classification utilizing an artificial neural network (ANN), which achieved an unprecedented peak accuracy of 92.48%, establishing a new performance benchmark for this specific dataset.

Artificial intelligence increasingly informs consequential decisions, yet high predictive performance alone does not establish that an intelligent system is worthy of trust. This systematic review examines trustworthy AI as a multidimensional, socio-technical, and lifecycle-dependent research domain. It synthesises selected peer-reviewed journal evidence addressing transparency, interpretability, explainability, robustness, safety, privacy, security, fairness, accountability, governance, and meaningful human oversight. The review applies predefined questions, eligibility principles, trustworthiness-dimension mapping, claim-level evidence extraction, and cross-dimensional synthesis without undertaking meta-analysis or inventing quantitative screening results. The evidence indicates that transparency and explanation must be assessed in relation to fidelity, stability, completeness, usefulness, stakeholder needs, contestability, recourse, and the practical authority of human overseers. Robustness and safety require evaluation under adversarial conditions, uncertainty, context shift, system failure, and residual risk, while privacy-preserving approaches must still be examined for leakage, cybersecurity vulnerabilities, and governance limitations. Fairness depends on context-sensitive definitions, bias identification, distributional consequences, documentation, responsibility allocation, auditability, and institutional mechanisms for correction. Cross-dimensional analysis reveals persistent tensions between transparency and privacy, explainability and security, automation and human control, and performance and fairness. Important gaps remain in multidimensional evaluation, realistic human-oversight studies, explanation-usefulness validation, lifecycle monitoring, accountability operationalisation, stakeholder participation, and foundation-model assurance. The review contributes an integrated evidence synthesis showing that AI worthiness of trust should be supported through coordinated technical validation, human-centred assessment, institutional governance, and continuing monitoring rather than inferred from any isolated property.

Explainability has become central to discussions of trustworthy and responsible intelligent systems, yet it is frequently reduced to the technical production of information about model structure, features, rules, or predictions. This critical review examines why opening a computational black box is not equivalent to producing an explanation that people can understand or use responsibly. The review synthesises conceptual, technical, human-centred, decisional, and accountability-oriented scholarship to evaluate the shift from model interpretability toward meaningful explanation. It distinguishes technical properties such as transparency, fidelity, plausibility, completeness, stability, and local or global scope from human outcomes such as comprehension, cognitive accessibility, mental-model development, decision usefulness, calibrated trust, and appropriate reliance. It further examines how explanations may produce overreliance, underreliance, automation bias, false reassurance, cognitive overload, or explanation mismatch when their content and presentation are not aligned with users, tasks, uncertainty, and decision consequences. From an accountability perspective, explanations may contribute to auditability, traceability, contestability, and recourse, but they cannot independently assign responsibility, provide institutional authority, or guarantee effective remedies. The main contribution is a human-centred critical synthesis that treats explainability as a socio-technical achievement rather than a model output. Meaningful explanation requires technical credibility, cognitive accessibility, contextual relevance, decision-sensitive evaluation, uncertainty communication, opportunities for challenge, and governance mechanisms capable of acting on explanatory evidence. The review therefore rejects universal explanation-sufficiency claims and identifies conditions under which explanations may assist understanding and responsible decision use without being treated as proof of trustworthiness, accountability, fairness, safety, or deployment readiness.

Intelligent systems increasingly influence decisions, allocate attention, generate recommendations, and mediate interactions across technical, organizational, and social environments. Yet trustworthiness is often approached through isolated interventions, such as improving model accuracy, adding post-hoc explanations, documenting datasets, or assigning nominal human oversight. This article proposes a trustworthy-by-design conceptual framework that treats trustworthiness as a system-level property requiring coordinated design and evaluation across the full intelligent-system lifecycle. The framework begins with system-purpose and risk-context definition, followed by stakeholder and responsibility mapping, data-integrity assessment, provenance and lineage review, data documentation and governance, model-reliability evaluation, robustness and uncertainty assessment, and bias and fairness review where relevant. It then connects explanation quality with human interpretability and decision usefulness, rather than assuming that technically generated explanations will necessarily support understanding. Human oversight, meaningful control, contestability, recourse, institutional accountability, traceability, auditability, monitoring, and feedback updating form additional interdependent layers. The framework’s principal contribution is to clarify how technical evidence must be connected to human decision conditions and institutional responsibility without treating any individual safeguard as sufficient. Data provenance does not establish data appropriateness, robustness does not prove safety, explanations do not guarantee interpretability, human presence does not ensure meaningful control, and governance arrangements do not independently establish accountability. The framework may support conceptual analysis, assurance-oriented evaluation, and lifecycle governance planning, but requires empirical validation, human-centered assessment, domain-specific interpretation, and continuing institutional review.

Explainable artificial intelligence increasingly offers diverse ways to communicate how intelligent systems operate, yet explanation selection is still frequently driven by method availability or model compatibility rather than stakeholder need. This article develops a stakeholder-oriented taxonomy for determining which explanation type may be appropriate for whom, at which decision moment, and for what purpose. Explanation is treated as a multiform system capability encompassing different scopes, reasoning functions, presentation modes, interaction patterns, and quality requirements. The taxonomy distinguishes global explanations concerned with model-level behaviour from local explanations directed at individual predictions, while clarifying the roles and limitations of feature attribution, saliency, surrogate models, rule extraction, transparency, fidelity, stability, and completeness. It further differentiates counterfactual, contrastive, and causal explanations according to their contributions to alternative reasoning, actionability, contestability, recourse, and accountability. Example-based, case-based, prototype, rule-based, natural-language, interactive, dialogue-based, adaptive, and personalized explanations are positioned according to user expertise, cognitive accessibility, task demands, and interaction context. The proposed taxonomy connects explanation families with stakeholder roles, pre-decision, during-decision, and post-decision timing, explanation purpose, decision consequence, uncertainty, trust calibration, human oversight, and accountability requirements. Its central contribution is a structured selection logic that moves beyond technical method categories toward stakeholder–purpose–timing–risk alignment. The taxonomy can guide explanation design and evaluation but does not independently establish understanding, fairness, recourse, accountability, legal sufficiency, trustworthiness, or deployment readiness. Stakeholder-specific and decision-context-specific validation therefore remains necessary.

Artificial intelligence decisions are often described as explainable when a system can produce reasons, feature contributions, counterfactuals, or other interpretive outputs. Such outputs, however, do not by themselves make an institutional decision answerable. This article develops a conceptual explainable AI governance framework that treats explainability as an organizational obligation connected to transparency duties, system documentation, evidence preservation, auditability, responsibility allocation, contestability, human review, and redress. The framework distinguishes technical explanation from the governance capacities required to identify who must justify a decision, what information must be recorded, how evidence can be examined, how affected stakeholders may challenge an outcome, and whether an organization can correct errors or remediate consequences. Transparency and documentation are organized around system, model, data, dataset, and decision records, together with data, model, and decision provenance. Auditability is connected to traceability, lifecycle records, risk documentation, monitoring, incident response, and clearly allocated organizational responsibility. Contestability is connected to decision justification, appeal and complaint pathways, competent and sufficiently independent human review, escalation, correction, recourse, and redress capacity. The framework also identifies governance gaps that arise when these functions are separated across the AI lifecycle, including documentation, evidence-trail, accountability, human-review, monitoring, and remediation gaps. Its principal contribution is an answerability-oriented governance architecture that links explanation to institutional evidence and response mechanisms while preserving explicit boundaries: transparency does not independently establish accountability, auditability does not prove responsibility, human involvement does not guarantee meaningful review, and a conceptual governance framework does not establish legal compliance, procedural fairness, ethical sufficiency, or deployment readiness.

Trustworthiness is frequently treated as a label attached to an intelligent system once selected technical, ethical, or governance requirements have been addressed. Such a binary interpretation is inadequate because system properties, operating conditions, organizational responsibilities, stakeholder expectations, and available assurance evidence change throughout the lifecycle. This article proposes a conceptual trustworthiness maturity model that interprets trustworthiness as an evolutionary system capability rather than a fixed attribute, certification outcome, or deployment decision. The model organizes progression across six levels: unstructured or ad hoc practices; foundational transparency and documentation; managed reliability, explainability, and human oversight; defined governance integration and accountability; continuous assurance and lifecycle monitoring; and adaptive trustworthiness learning and system updating. Foundational maturity requires accessible system, model, data, dataset, and decision documentation together with provenance, traceability, auditability, and maintained evidence trails. Higher capability requires context-sensitive evidence of reliability, validity, robustness, uncertainty, calibration, explanation quality, and meaningful human intervention. Advanced maturity additionally depends on responsibility mapping, organizational and institutional accountability, lifecycle governance, continuous monitoring, drift review, incident learning, assurance review, and controlled feedback updating. The principal contribution is an integrated maturity structure that connects technical, human, organizational, and lifecycle capabilities while retaining explicit boundaries against overclaiming. Maturity levels may help structure assessment planning, identify capability gaps, and clarify development priorities, but they do not independently prove trustworthiness, safety, ethical sufficiency, legal compliance, certification eligibility, or deployment readiness. The proposed model therefore supports conceptual maturity assessment and future empirical validation rather than operational prescription.

Foundation models have expanded the scope of artificial intelligence by supporting language generation, multimodal reasoning, knowledge-intensive assistance, decision support, and adaptable downstream applications. Their broad capabilities, however, do not establish that their outputs, explanations, internal processes, or deployment arrangements are trustworthy. This scoping review maps the research landscape concerning foundation-model trustworthiness across interpretability, explanation capabilities, hallucination, uncertainty, reliability, alignment, evaluation, and human oversight. The review uses structured search concepts, eligibility principles, evidence extraction, and dimension-based mapping to examine how peer-reviewed scholarship characterizes foundation models as general-purpose intelligent systems and how trustworthiness evidence is distributed across technical, human-centered, organizational, and governance domains. The synthesis indicates that interpretability methods can expose selected behavioral or representational features but do not independently demonstrate explanation fidelity, usefulness, or safety. Hallucination and factuality research identifies persistent risks involving fabrication, unsupported claims, weak grounding, and persuasive misinformation, while uncertainty methods remain constrained by calibration, context shift, and user-communication challenges. Alignment and post-training techniques may shape preferred behaviors but retain risks associated with proxy objectives, preference representation, and residual failure. Benchmarking, human evaluation, stress testing, and oversight provide partial evidence, yet their value depends on construct validity, domain relevance, intervention capacity, accountability, and lifecycle monitoring. The research landscape is therefore fragmented across disciplines and frequently model-centred rather than socio-technical. Key priorities include stronger foundation-model interpretability validation, consistent hallucination measurement, usable uncertainty communication, externally valid evaluation, meaningful oversight testing, multimodal analysis, and lifecycle governance. The review contributes an integrated evidence map while emphasizing that no single trustworthiness dimension certifies safety, reliability, accountability, or deployment readiness.

Explainable artificial intelligence has generated extensive technical and human-centered evaluation research, yet substantial uncertainty remains about what explanation evaluations actually demonstrate. This umbrella review synthesizes review-level evidence concerning the evaluation of fidelity, faithfulness, stability, robustness, comprehensibility, cognitive accessibility, decision usefulness, human trust, trust calibration, and reliance. Rather than reviewing explanation algorithms as methods, the article examines the constructs, metrics, study designs, and validity assumptions through which explanation quality has been assessed. The synthesis distinguishes technical correspondence with model behavior from human judgments of plausibility, understanding, usefulness, and satisfaction. It further separates human trust from system trustworthiness and behavioral reliance from beneficial decision support. Review-level evidence indicates that technical metrics are heterogeneous, terminology is inconsistently operationalized, and stable or plausible explanations are not necessarily faithful. Human-centered evaluations frequently rely on self-reported understanding, satisfaction, or trust, while objective comprehension, cognitive load, decision quality, appropriate reliance, overreliance, and underreliance receive less consistent assessment. Evidence linking technical explanation properties to downstream human and institutional outcomes remains limited, particularly across stakeholder groups, realistic decision contexts, repeated use, and changing system conditions. The umbrella synthesis therefore rejects single-metric interpretations of explanation quality and organizes evaluation as a multidimensional evidence problem involving technical validity, cognitive accessibility, decision consequences, calibrated trust, reliance behavior, accountability, and contestability. Its principal contribution is a consolidated evaluation agenda that clarifies what existing measures may support, what they do not establish, and which methodological connections require stronger user-centered, task-specific, longitudinal, and socio-technical validation.

Explainable artificial intelligence is often presented as a means of making intelligent systems more understandable, trustworthy, and usable, yet explanation availability does not reliably produce adoption, appropriate reliance, accountability, or improved decisions. This realist review examines why explainable AI is adopted, ignored, rejected, discontinued, overused, underused, or misused across organizational, professional, technical, and decision settings. It applies context–mechanism–outcome reasoning to interpret how explanation design interacts with workflow integration, user and domain expertise, decision risk, organizational incentives, governance arrangements, and human-oversight conditions. The synthesis focuses on mechanisms including perceived transparency, cognitive accessibility, mental-model formation, perceived usefulness, trust calibration, accountability confidence, contestability awareness, algorithm aversion, algorithm appreciation, automation bias, and explanation overload. These mechanisms may support selective reliance and meaningful challenge, but they may also generate false reassurance, superficial compliance, defensive rejection, or uncritical dependence. The review distinguishes adoption from acceptance, trust from trustworthiness, reliance from appropriate reliance, and explanation usefulness from decision quality. It identifies plausible configurations in which professional expertise, uncertainty communication, traceability, feedback, and contestability support careful use, while poor workflow fit, limited accessibility, weak governance, and automation pressure suppress scrutiny or encourage misuse. The main contribution is a bounded realist synthesis and design agenda for evaluating explainable AI through context-sensitive human responses and decision outcomes rather than explanation availability alone. The configurations are explanatory propositions requiring human-centered, organizational, and decision-context validation; they are not universal causal laws, implementation protocols, or evidence of deployment readiness.

Explainable artificial intelligence increasingly seeks to provide more than visualisations of feature importance or simplified descriptions of model outputs. For intelligent systems used in consequential settings, machine reasoning should be open to questioning, inspection, contextualisation, correction, and evaluation under uncertainty. This integrative review examines emerging approaches that may support this broader objective. It connects causal explainability and counterfactual reasoning with neuro-symbolic intelligence, knowledge-grounded inference, uncertainty-aware explanation, reliability evaluation, and human-centred interpretability. The review uses a structured integrative approach to compare how these paradigms represent causes, alternatives, rules, concepts, knowledge sources, inferential steps, confidence, and limitations. Causal and counterfactual approaches can support questions about mechanisms, contrasts, interventions, actionability, and recourse, but their conclusions depend on explicit assumptions, feasible alternatives, and appropriate causal evidence. Neuro-symbolic and knowledge-grounded approaches can make concepts, relations, rules, provenance, and intermediate reasoning more traceable, although symbolic structure or retrieved knowledge does not automatically ensure faithful, valid, or truthful explanation. Uncertainty-aware approaches can qualify predictions and explanations through uncertainty characterisation, calibration, confidence communication, robustness assessment, and human interpretation, but uncertainty estimates alone do not establish reliability. The synthesis indicates that no individual paradigm is sufficient. More defensible machine reasoning requires complementary evaluation of explanation fidelity, causal plausibility, knowledge validity, reasoning traceability, uncertainty communication, cognitive accessibility, contestability, and lifecycle accountability. Important research opportunities concern causal validation, feasible recourse, provenance governance, explanation uncertainty, hybrid-system evaluation, human-centred testing, and cross-paradigm assessment. The review contributes an integrated account of how machine reasoning may become not merely visible, but meaningfully open to question.

Foundation models are increasingly embedded in information access, content production, analysis, communication, and decision-support processes. Their broad capabilities have intensified interest in alignment, yet alignment alone cannot establish responsible use. A model may appear responsive to instructions or preferences while remaining opaque, weakly grounded, unreliable under changing conditions, overconfident, difficult to contest, or insufficiently governed. This article proposes a conceptual governance and technical safeguard framework for responsible foundation models. The framework treats foundation models as components of decision infrastructure situated within socio-technical and institutional environments rather than as isolated predictive tools. It integrates eight interdependent areas: model and alignment safeguards; explainability and interpretability; knowledge grounding and provenance; reliability, uncertainty, and hallucination containment; human control and escalation; institutional accountability; auditability and documentation; and lifecycle monitoring and governance updating. Explainability requirements are differentiated from interpretability, transparency, explanation usefulness, and human understanding, while knowledge grounding is separated from factuality, truthfulness, source quality, and provenance. Reliability is addressed through calibration, robustness assessment, boundary testing, uncertainty communication, failure containment, and continuing monitoring. Human oversight is defined through authority, competence, review capacity, escalation pathways, contestability, and recourse rather than nominal human presence. Institutional responsibility is connected to role allocation, evidence trails, documentation, review authority, and consequences. The principal contribution is an integrated framework that clarifies how technical and governance safeguards may support responsible foundation-model assessment without treating alignment, explanation, grounding, testing, oversight, or auditing as independently sufficient. The framework requires empirical validation, domain-specific evaluation, human-centred assessment, institutional governance, and residual-risk management before it can inform consequential operational decisions.

Generative artificial intelligence can produce fluent, context-sensitive, and apparently authoritative responses even when relevant knowledge is incomplete, evidence is unavailable, or generated claims are false. This creates an epistemic safety problem concerning not only what a system produces, but also what it appears to know, how its claims are supported, how uncertainty is expressed, and when answering should be limited. This article develops a conceptual safety framework organized around four operational principles: ground, qualify, verify, and abstain. The framework treats hallucination as a multidimensional risk involving fabrication, factuality failure, groundedness failure, unsupported claims, false confidence, source and context errors, reasoning errors, citation errors, overgeneralization, confabulation, speculative output, and incomplete evidence. Grounding links material claims to relevant and sufficiently reliable evidence; qualification communicates uncertainty, assumptions, scope, and evidential limitations; verification examines claims, sources, citations, and inferential consistency; and abstention enables refusal, deferral, escalation, or constrained output when evidence or system competence is insufficient. Human review, meaningful oversight, decision boundaries, failure containment, incident monitoring, and lifecycle updating operate across these principles rather than functioning as final guarantees. The main contribution is an integrated generative AI safety framework that connects output-level hallucination control with source quality, knowledge verification, uncertainty communication, review quality, response limitation, and downstream risk containment. The framework is intended to structure domain-specific assessment and future research. It does not independently establish factual reliability, ethical sufficiency, legal compliance, operational safety, certification, or deployment readiness.

Agentic artificial intelligence extends beyond conventional model-output generation by interpreting objectives, decomposing tasks, maintaining contextual information, selecting tools, interacting with external environments, and executing sequences of actions that may create downstream consequences. These capabilities complicate the question of who controls an agent and expose limitations in governance approaches based only on alignment, instruction following, output review, or nominal human supervision. This article develops a conceptual governance architecture for trustworthy agentic AI. The approach treats agentic AI as an autonomous decision system embedded within a wider socio-technical environment and organizes governance around explicit control points rather than a single supervisory mechanism. It distinguishes governance requirements for goal formation, goal interpretation, instruction hierarchy, task boundaries, tool access, memory retention, environmental interaction, action verification, failure containment, human approval, override capacity, responsibility allocation, auditability, and lifecycle review. The architecture proposes that consequential actions should be connected to authorized goals, bounded permissions, traceable memory and data inputs, pre-action checks, post-action evaluation, escalation pathways, and accountable human or institutional actors. It also establishes explicit sufficiency boundaries: goal alignment does not guarantee appropriate behavior; tool restrictions do not eliminate harmful use; memory controls do not guarantee reliable context; verification does not certify safe action; and human override does not independently establish meaningful control. The principal contribution is an integrated architecture-level account of how technical, human, and organizational controls may be connected across the agent lifecycle. The architecture supports conceptual analysis, governance planning, and future empirical evaluation, but does not establish legal compliance, operational safety, ethical sufficiency, institutional accountability, or deployment readiness.

Multimodal intelligent systems combine evidence from text, vision, audio, sensors, and structured data, but their explanations often remain fragmented into separate saliency maps, feature scores, textual rationales, attention visualisations, or confidence indicators. Such outputs may describe isolated aspects of a model while obscuring how modality-specific evidence was represented, transformed, fused, and communicated. This article develops a conceptual multimodal explainability architecture organised around the principle of following evidence across modalities. The approach begins by distinguishing the evidential and representational characteristics of textual, visual, acoustic, temporal-sensor, and structured inputs. It then identifies attribution difficulties created by early, intermediate, late, and hybrid fusion, including modality dominance, missing evidence, conflicting inputs, representation alignment, and uncertainty about joint contributions. Evidence traceability is treated as an end-to-end requirement connecting input provenance, modality-specific representation, fusion operations, attribution processes, consistency assessment, uncertainty communication, and decision explanation. The architecture further incorporates human interpretation because technically detailed multimodal explanations may remain cognitively inaccessible, visually persuasive, incomplete, or poorly calibrated to user expertise and decision context. Its central contribution is a layered framework linking modality inventory, evidence classification, representation mapping, fusion attribution, provenance, consistency review, human-centred presentation, evaluation, and feedback updating. The architecture is intended to structure research and design reasoning rather than prescribe a particular model, explanation technique, evaluation metric, or deployment procedure. It does not establish that an explanation is faithful, that users understand it, that reliance is appropriate, or that the underlying intelligent system is trustworthy or ready for operational use.