Search Results (88)

Background: Artificial intelligence (AI) adoption is increasingly reported as a priority for organizations, yet they face a growing, fragmented body of evidence concerning the factors that influence successful AI integration. Method: To identify the relevant factors for organizational AI adoption, we conducted a systematic literature review (SLR) following PRISMA guidelines, which yielded 37 quantitative empirical studies. From these studies we extracted 1229 paper-item instances, of which 810 were retained after applying structured exclusion criteria to develop a domain model relevant to organizational AI adoption. The model’s content validity was assessed and supported through expert feedback using the Content Validity Index (CVI) methodology. Results: We organized 24 subclusters into nine main clusters across the three dimensions Technology (Enablers, Usability, Trust), Organization (Leadership, People, Process), and Environment (Market, Regulatory, Partner). Our analysis suggests that workforce skills, perceived intelligence, and resources are among the most frequently studied and positively associated antecedents of AI adoption, and that constructs related to AI explainability and control (human-in-the-loop oversight) have received little research attention and remain underrepresented despite growing regulatory requirements such as the EU AI Act. Conclusions: The resulting domain model provides an empirically grounded classification of organizational AI adoption factors and can serve as a foundation for future measurement instruments. Full article

Ensemble learning has been widely adopted in fraud detection; however, conventional ensemble strategies rely on uniform or performance-based weighting schemes that treat explainability as a post hoc annotation rather than an architectural component. This study addresses the research goal of whether SHAP attribution values can serve as a principled, instance-specific weighting mechanism within an ensemble, thereby embedding interpretability directly into the aggregation process. A SHAP-Weighted Ensemble (SWE) framework is proposed in which the L2 norm of each base model’s SHAP attribution vector, computed at prediction time, is used to derive instance-specific voting weights via Softmax normalization. Three linear base learners (logistic regression, robust LR, calibrated linear SVM) are combined, with LinearSHAP providing exact attribution values. A comprehensive evaluation protocol was applied on a real-world vehicle insurance claims dataset, including bootstrap 95% confidence intervals, McNemar’s test, a three-way ablation study comparing equal weighting, SWE, and validation-AUC weighting, F1-optimal threshold selection, expected calibration error, and cost-sensitive evaluation under asymmetric misclassification costs. The central finding is that SWE achieves performance statistically comparable to both simpler baselines across all evaluated metrics (ROC-AUC = 0.774, 95% CI [0.681, 0.862]; F1 = 0.679, 95% CI [0.569, 0.774]; McNemar p = 1.000), while producing a transparent, per-claim weighting trace that equal-weight voting cannot provide. A KernelSHAP influence analysis conducted directly on the SWE confirms that SHAP-derived weights are substantially aligned with actual model influence ratios (LR: 1.05×, LR_R: 1.05×, SVM: 0.81×), validating the weighting mechanism empirically. An exploratory analysis of a seven-model equal-weight diagnostic ensemble reveals a negative correlation (r = −0.721, p = 0.067) between individual model performance and ensemble influence; a theoretically coherent finding that does not reach statistical significance at conventional thresholds. The primary contribution of SWE is architectural and interpretability-driven: it produces an auditable, instance-level model-weighting mechanism grounded in SHAP attribution theory, supporting regulatory accountability under GDPR Article 22 and the EU AI Act. Full article

Artificial Intelligence (AI) shows significant potential across healthcare domains, including advanced diagnostics, clinical decision support, and personalized medicine. Despite these advancements, the opaque ‘black box’ nature of complex AI models necessitates the application of Explainable Artificial Intelligence (XAI) to ensure trust, accountability, interpretability, and regulatory compliance. This study systematically reviews 76 studies published between 2020 and 2025 that have used XAI in healthcare. Our findings show that XAI models such as SHAP and LIME are predominantly used for structured data applications, such as electronic health records, while other XAI models, such as Grad-CAM and Layer-wise Relevance Propagation (LRP), are mainly used in medical imaging. This study specifically investigates evaluation metrics for operationalizing explainability, including faithfulness, trustworthiness, and regulatory compliance, which distinguishes it from prior descriptive reviews. Our analysis shows that while XAI significantly enhances clinician trust, thorough explanation remains heterogeneous and largely confined to controlled settings and the employed benchmark datasets. Critical barriers to clinical adoption include inconsistent interpretability across data modalities and the lack of standardized evaluation frameworks. Existing XAI techniques often do not correspond with strict regulatory standards such as the EU AI Act, Food and Drug Administration (FDA) guidelines, and the Health Insurance Portability and Accountability Act (HIPAA). This review argues for the urgent standardization of XAI validation and the development of human-centered designs to move beyond algorithmic transparency toward reliable real-world hospital integration. Full article

Autonomous artificial intelligence (AI) systems increasingly participate in decision-making processes that affect individuals, markets, and public administration. Their growing autonomy complicates the attribution of legal responsibility, particularly within regulatory frameworks that were designed around identifiable human actors and relatively stable products. Although European Union instruments such as the GDPR, the AI Act, and the revised Product Liability Directive address specific dimensions of risk and compliance, they do not fully resolve how responsibility should be allocated across the lifecycle of complex AI systems. The difficulty does not lie so much in the absence of legal rules. Rather, it reflects the structural tension between traditional liability models and the distributed architecture of contemporary AI development and deployment. By examining how existing EU regulatory instruments interact, the paper identifies fragmentation in responsibility allocation that may weaken institutional accountability. It then proposes a tiered model of legal responsibility based on meaningful control at different stages of system design, deployment, and operational oversight. Rather than introducing new forms of legal personhood, the model seeks to clarify how existing doctrines can be interpreted and coordinated in order to maintain regulatory coherence and socially intelligible accountability in digitally mediated environments. The model allocates responsibility according to meaningful control within distributed systems, offering a structurally coherent alternative for EU governance. Full article

Reducing building energy consumption is central to EU climate-neutrality targets and to sustainable development goals: buildings account for around 40% of EU final energy consumption, placing Building Energy Management Systems (BEMS) at the intersection of the European Green Deal and the EU Artificial Intelligence Act. A scoping review following PRISMA-ScR guidelines charted 61 Machine Learning (ML) for BEMS papers (2020–2026) across three sub-domains (load forecasting and energy monitoring, HVAC control, and demand response), using a nine-point Technology Readiness Level (TRL) rubric and three Trustworthy AI (TAI) dimensions (Privacy & Data Governance, Robustness, and Transparency). The review finds that 90.2% of papers remain at the development stage (TRL 4–6), with no multi-site production deployment documented. TAI coverage is heterogeneous at publication level: transparency is addressed in only 3 of 61 papers (4.9%), and privacy provisions (the best-covered ALTAI dimension) are concentrated in demand-response papers (9 of 17, 52.9%), largely via Federated Learning (6 of 9 privacy-tagged papers). A three-level EU AI Act risk classification identifies 23 borderline-candidacy papers (37.7%), predominantly Reinforcement Learning-based HVAC control systems, whose high-risk proximity cannot be resolved at abstract level; explicit compliance engagement is absent from all 61 mapped sources, including the 22 papers published after the Act entered into force in August 2024. The findings document adouble readiness gap: a TRL ceiling co-located with limited documented engagement with TAI obligations and EU AI Act compliance at publication level. Closing this gap is necessary before AI-driven building energy management can be deployed at scale under EU governance requirements. Full article

Artificial intelligence (AI) is becoming increasingly embedded in the digital infrastructure of local government, creating new opportunities to improve public sector productivity while also influencing systemic risk and organisational resilience across interconnected public systems. As municipalities adopt AI to automate, support, and transform administrative processes, organisational performance becomes more dependent on the reliability of algorithms, the quality of data, effective governance, and coordination among public institutions. These growing interconnections create new vulnerabilities that can spread across public service networks, yet evidence on the productivity, risk, and resilience implications of AI adoption remains fragmented and dispersed across different fields of research. This study develops an integrative conceptual framework that examines the relationship between AI adoption, public sector productivity, systemic risk, and organisational resilience within interconnected sociotechnical systems. Drawing on insights from productivity economics, systems theory, and public governance, the framework positions total factor productivity (TFP) within a broader public value and risk governance perspective. Using the PRISMA 2020 methodology, the study systematically reviews 68 peer reviewed empirical studies published between 2015 and 2025, assessing productivity outcomes, methodological quality, effect sizes, and contextual factors relevant to local government and networked public administration. The findings show that productivity gains associated with AI are strongly influenced by organisational readiness, including digital maturity, workforce capabilities, governance quality, and institutional coordination. While AI has the potential to improve operational efficiency and strengthen adaptive capacity, inadequate readiness can increase systemic risks arising from algorithmic opacity, cybersecurity challenges, data dependence, coordination failures, and disruptions that may spread across interconnected administrative systems. The review also highlights that resilience depends on the ability of public organisations to anticipate, absorb, adapt to, and recover from AI-related disruptions while maintaining the continuity and quality of public services. The study contributes to theory by integrating perspectives from productivity economics, public administration, and systemic risk within a sociotechnical systems framework. It contributes empirically through a comprehensive synthesis of evidence on AI and public sector productivity and methodologically through the application of transparent PRISMA 2020 review procedures. From a practical perspective, the study offers a conceptual measurement framework and policy guidance for municipal decision makers seeking to improve productivity while strengthening resilience and reducing systemic risks in increasingly interconnected public governance systems. Full article

Integrated care is increasingly shaped by digital infrastructures, data governance, and AI-enabled analytics, yet the relevant literature remains fragmented across health-services research, digital health, and machine learning. This article reports a scoping review, conducted in line with PRISMA-ScR guidance, that maps how integrated care models have evolved conceptually, what digital and AI-enabled infrastructures support them, how their clinical, economic, and equity impacts can be evaluated, and what current implementations imply for sustainable scaling. We searched PubMed, Scopus, Semantic Scholar, and Crossref (retrieval date 31 October 2025; forward screening to 31 March 2026) and added grey literature from named policy bodies. The searches identified 15,189 records, reducing to 11,789 after intra- and cross-source deduplication and grey-literature integration; 620 full texts were assessed and 192 were included in the synthesis. Four domains were synthesised: conceptual foundations of integrated care, AI and multimodal analytics, implementation barriers, and digital-governance foundations. We chart the field using a Type I–V maturity scheme (disease, cohort, whole-system, digital-integrated, learning), benchmarked against the Rainbow, MacColl, EMRAM/AMAM, and NHS ICS models. Most deployments cluster at digitally integrated but only weakly adaptive Type IV; recurrent failure modes—temporal blind spots, maintenance debt, semantic drift, and governance gaps—block progression to Type V, and high-profile clinical-AI failures illustrate the cost of attempting Type V analytics on Type IV-or-worse infrastructure. A walk through nine world regions maps each to its current Type I–V position and shows that organisational and payment integration—not digital sophistication alone—is currently the dominant driver of progress. The COMFORTage Integrated Care Model Library is positioned as a workflow of AI agents orchestrating predictive, preventive, and personalised care across the integrated-care lifecycle rather than as a single federated-learning programme. The review positions AI-enabled integrated care less as a finished model than as an emerging design space requiring longitudinal data assets, stewarded model lifecycles, accountable governance, and outcome-based contracting for clinically useful, equitable, and trustworthy learning systems. Full article

Artificial Intelligence Chatbots based on Large Language Models are becoming important cornerstones in education, professional, and personal contexts. Users frequently disclose sensitive information without the necessary awareness of how their data is processed, stored, and later used, creating risks of unintended exposure and violation of data protection regulations such as GDPR and the EU AI Act. In this paper, we present a dynamic privacy-by-design framework that introduces a novel privacy-oriented middleware layer between the users and the AI Chatbots. Our framework intercepts user prompts, detects and sanitizes sensitive information, and enables explicit user control over data retention and disclosure through our dedicated Information Manager Dashboard. To develop our framework around real-world needs, we conducted a survey with 83 participants investigating privacy concerns, regulatory awareness, and preferences for transparency and control in conversational AI. Our results indicate that better and more transparent privacy safeguards can be achieved without significantly compromising usability or performance, supporting the development of trustworthy and user-centric AI chatbots. We also evaluate our solution against other state-of-the-art implementations and relevant metrics. Full article

Human-Centered Artificial Intelligence (HCAI) has emerged as a promising paradigm to increase transparency, usability, and trust in AI-driven Decision Support Systems (DSS). However, existing research lacks technically detailed accounts of how HCAI principles can be operationalized, implemented, and empirically validated in real decision environments. This study proposes a technically grounded HCAI-oriented DSS framework and presents a concrete prototype implemented in two high-stakes domains: clinical decision support and financial risk assessment. The architecture integrates interpretable machine learning models, SHAP-based explanations, structured user-feedback loops, and governance mechanisms aligned with the EU Trustworthy AI Guidelines. We trained and evaluated domain-specific models using publicly available medical and financial datasets, describing all data preprocessing, model selection, and hyperparameter settings to ensure reproducibility. An empirical study involving 30 domain experts (15 clinicians, 15 financial analysts) compared the HCAI-DSS with a functionally identical black-box DSS. Statistical analyses (paired t-tests with 95% confidence intervals and Cohen’s d) revealed that the HCAI-DSS significantly improved trust (d = 1.23), transparency and understanding (+1.76 mean difference), usability (SUS difference = +15.4), and decision accuracy (+10.2%), without a significant increase in decision time (p = 0.08). Qualitative feedback further demonstrated that explanations, control, and human-in-the-loop features increased confidence and reduced uncertainty. The results provide empirical evidence that HCAI principles tangibly enhance DSS effectiveness and user acceptance. The study contributes (1) a reproducible technical implementation, (2) a validated HCAI-DSS architecture, and (3) multi-domain evidence of improved decision quality. These findings support sustainable and trustworthy AI adoption across sectors and align with emerging regulatory frameworks such as the EU AI Act. Full article

Large Language Models (LLMs) are rapidly transitioning from research concepts to transformative artificial intelligence components within the power and energy domain. Their ability to fuse diverse data, spanning SCADA logs, real-time sensor readings, and regulatory documentation enables unprecedented capabilities in forecasting, operator decision support, anomaly detection, and wide-area situational awareness for future intelligent grids. However, the integration of LLMs into safety-critical and highly regulated power systems introduces a convergence of novel and severe security risks. Beyond exhibiting model-intrinsic vulnerabilities like hallucination, prompt injection, and data poisoning, these models are susceptible to system-level threats that could compromise grid stability, distort energy market operations, or facilitate the leakage of sensitive operational data. Moreover, integrating LLM workloads into cloud or hybrid architectures necessitates strict compliance with critical standards and emerging governance frameworks like the EU AI Act. While existing surveys address AI security in power systems, general LLM security, and AI in smart grids separately, this paper bridges these threads by providing a unified treatment of LLM-specific risks, power-system deployment constraints, and emerging governance frameworks—a combination not covered in prior surveys. We provide a systematic taxonomy of risks across five dimensions: cybersecurity, privacy, robustness, explainability, and governance. We synthesize technological advances, clarify the complex interplay between LLM failure modes and grid security, and propose a forward-looking research agenda to guide future investigation. This work aims to be an indispensable resource for researchers, utility operators, and policymakers in designing resilient, trustworthy, and compliant AI-enabled energy infrastructures. Full article

As software systems increasingly combine machine learning, deep learning, and generative AI components with classical deterministic logic, the systematic detection of AI-based algorithmic elements in application code is becoming essential for software audit, compliance with the EU AI Act (Regulation (EU) 2024/1689), and quality assurance. This paper introduces Quality-Informed Neural Networks (QINN), an architecture in which the structured knowledge encoded in the Quality Function Deployment (QFD) House of Quality is embedded into the network topology and weight initialisation through QFD-derived binary structural masks and knowledge-calibrated initialisation—in direct analogy with Physics-Informed Neural Networks (PINNs). The QFD relationship matrices act as structural priors that constrain the hypothesis space toward quality-consistent solutions by enforcing domain-expert-validated sparsity on network connectivity, while an optional QFD-regularised loss term provides an additional soft constraint on the learned weight structure. As a proof of concept, QINN is instantiated in its masked-architecture configuration for the binary classification of software repositories as AI-enabled or classical. On the AIC-199 proof-of-concept dataset, the proposed QINN attains a cross-validated AUC of 99.47% (±1.18%), recall of 100.00% (±0.00%), and F1-score of 99.02% (±1.34%) under QFD-informed structural masking, outperforming a non-learned QFD scoring baseline by 37.37 percentage points in recall and exceeding a cross-validated Random Forest ensemble on AUC by 2.47 percentage points (W = 0, p < 0.05), while producing explanations at three QFD-grounded levels—feature salience, named Technical-Evidence activations, and per-criterion quality requirement scores—that align directly with the EU AI Act documentation obligations. Validation on larger, independently curated datasets and sensitivity analysis of the QFD elicitation process are identified as priorities for future work. A domain-general seven-phase application protocol is provided. Full article

The digital transformation of agriculture is rapidly turning the sector into a highly data-intensive domain. The European Union has responded with a broad horizontal framework encompassing the General Data Protection Regulation (GDPR), the Data Governance Act (DGA), the Data Act, the PSI Directive and the AI Act. However, this framework remains sector-neutral: it does not define ‘agricultural data’ as a legal category, nor does it explicitly recognize the specific position of farmers as data providers. This article pursues three objectives: (i) to map the EU legal and policy framework on data and AI as it applies to agriculture and identify regulatory gaps; (ii) to synthesize key concerns from the literature on agricultural data governance, with particular attention to the position of farmers and data spaces; and (iii) to develop an outline of a Spanish ‘Law on Agricultural Data and Digital Agricultural Services’ as an example of sectoral data legislation. The proposed Act—structured around a Preliminary Title and seven substantive Titles—would define agricultural data, recognize farmers as data providers, establish mandatory contractual protections, govern agricultural data spaces and cooperatives, introduce sector-adapted AI rules, address data sovereignty, and set up an institutional framework and graduated sanctions. The analysis argues that sectoral data law can complement EU horizontal rules, enhance legal certainty, and empower farmers without fragmenting the internal market. The article employs a doctrinal legal analysis and normative design-oriented methodology, drawing on secondary literature, policy documents, and EU and Spanish law; it does not rely on original empirical fieldwork. Full article

Regulatory sandboxes have evolved from specialised FinTech tools into broader mechanisms of regulatory experimentation. As artificial intelligence (AI) applications become embedded in credit decisioning, payment-fraud detection, identity verification, crypto-asset compliance, customer-facing advice and supervisory analytics, sandbox design increasingly affects how legal and institutional responsibility is allocated among regulators, participating firms, technology vendors and users. This article provides a comparative doctrinal and institutional analysis of accountability and liability in AI-related financial regulatory sandboxes. It clarifies the relevant AI modalities, distinguishes accountability (answerability and enforceability during sandbox participation) from liability (contractual, tort/product and regulatory/public law responsibility after harm), and maps framework-level safeguards across the European Union, the United Kingdom, Singapore, Norway and Hungary. The analysis does not seek to measure sandbox effectiveness empirically. Instead, it examines how publicly available legal and regulatory materials structure the allocation of duties before, during and after sandbox testing. The article shows that sandboxes generally do not operate as liability shields. Their legal significance lies in reallocating ex ante accountability duties—documentation, disclosure, monitoring, human oversight and exit planning—while preserving baseline liability rules. An Accountability and Liability Protocol is proposed to clarify roles, protect baseline consumer rights, support evidentiary traceability and connect sandbox learning to enforceable post-sandbox obligations. Full article

Operational Technology (OT) and Industrial Control Systems (ICSs) along the NATO-EU eastern flank face escalating hybrid threats, yet existing cyber-resilience metrics remain IT-centric, lacking OT-specific constraints and geopolitical exposure dimensions. This paper presents a Design Science Research contribution: the development and simulation-based feasibility demonstration of two interconnected artefacts. The first is the AI-enabled Cyber-Resilience Index (ACRI)—a composite 0–100 metric operationalized through 16 indicators across four domains (detection performance, operational continuity, governance maturity, supply-chain risk), aggregated as a three-term convex combination of capability domains with a linear subtractive supply-chain exposure penalty, weighted via AHP-based illustrative sector-reference profiles. The second is the Unified Compliance Framework (UCF), a structured R → C → E → SLO mapping linking 47 atomic regulatory requirements (NIS2, DORA, CER, AI Act, CRA) to standards (IEC 62443, ISO/IEC 27001) and auditable evidence artifacts, with a Continuous Assurance Loop operationalizing continuous control monitoring. Feasibility is demonstrated through digital twin simulation under three OT-representative threat scenarios (energy SCADA APT, railway supply-chain compromise, manufacturing ransomware). Results in simulated environments show ACRI improvement from Moderate-Risk baselines (45–61) to Adequate-Resilience thresholds (65–73); the proposed federated autoencoder–LSTM detector attains a composite $D_{perf}$ of 0.883 versus 0.510 for a static ±3$\sigma$ threshold baseline (a 73% relative improvement at the domain level). Sensitivity analysis confirms classification robustness ($\pm 7.3\%$ weight perturbation; coefficient of variation below 9.1% across 10,000 Monte Carlo iterations). Critical limitations are explicit: simulation-only evidence ($n=12$ scenario instances), illustrative (non-empirical) AHP weights, no operational field validation, and limited inferential statistical power. instances), illustrative (non-empirical) AHP weights, no operational field validation, and limited inferential statistical power. The contribution is positioned as a proof-of-concept design artifact establishing methodological foundations for OT-centric resilience assessment and compliance-to-engineering traceability, not as a field-validated operational system. Full article

The Court of Justice of the European Union’s Grand Chamber judgement in Public.Resource.Org v Commission (“Malamud”) raises fundamental questions about the relationship between EU law and the European standardisation system. By holding that harmonised standards referenced in the Official Journal must be accessible to the public free of charge, the Court reaffirmed the principle that individuals must be able to know the norms governing their legal rights and obligations. While widely welcomed as a victory for transparency and the Rule of Law, the judgement poses significant challenges to the longstanding governance and funding model of European standardisation, which relies heavily on copyright-based revenues generated by European Standardisation Organisations and national bodies. This article examines the practical and institutional consequences of the Malamud ruling and explores viable compliance pathways for the EU standardisation system. After outlining the role of harmonised standards within the New Legislative Framework and their growing importance in regulatory regimes such as the Artificial Intelligence Act, it analyses the judgement’s implications for access to law and the copyright status of standards. The article then evaluates a range of implementation models, including Commission-hosted publication, read-only access portals, licencing buyouts, and expanded use of common specifications. It argues that a Commission-hosted publication model, supported by revised funding arrangements, offers the most coherent pathway to reconcile open access with the continued functioning of the European standardisation infrastructure, and proposes corresponding reforms to Regulation 1025/2012. Full article