Frontiers in Artificial Intelligence最新文献

{"title":"Machine learning classification techniques for smart agriculture: a review of disease detection and crop recommendation.","authors":"M Umamaheswari, Kavitha D","doi":"10.3389/frai.2026.1878006","DOIUrl":"10.3389/frai.2026.1878006","url":null,"abstract":"<p><p>The adoption of Internet of Things (IoT) and Machine Learning (ML) in agriculture presents a revolutionary step toward efficient agricultural practices using data to address global issues surrounding food security. This review conducted a systematic search of the Scopus, IEEE Xplore, ScienceDirect, and SpringerLink databases for articles published between January 2023 and January 2026. It used keyword combinations like \"IoT and machine learning and agriculture\", \"crop recommendation\", \"plant disease detection\", and \"precision farming\". From the initial pool of articles found, 30 articles met the criteria for empirical methodology, quantitative performance reporting, and clear dataset descriptions. These were chosen for detailed analysis. It was discovered that the ensemble, as well as the traditional machine learning methods, such as Random Forest, Support Vector Machine (SVM), and Decision Trees, produce highly accurate predictions of the crop yields with an accuracy of 99.9%. Meanwhile, Convolutional Neural Networks (CNN) and ResNet-based deep learning models detect diseases with an accuracy above 98% while Gated Recurrent Units (GRU) and Long-Short Term Memory (LSTM) time-series models provide forecasts of weather conditions with an accuracy ranging from 94 to 98%, due to the complex nature of the environment dynamics. The application of IoT systems for precision agriculture produced a significant improvement in terms of resource utilization, resulting in 60-96% of water savings, 25% increased yield, and 36-38% reduced herbicide use. Although these advancements have been made, wider application is still hindered by high costs of implementation, lack of technological skills among farmers, data privacy concerns, issues of scalability within different agricultural settings, and consistent internet connection in rural areas. The current paper evaluates the use of IoT and ML in the agriculture sector from diverse data sources and studies to reveal the immense benefits offered by the two systems in enhancing farming activities.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1878006"},"PeriodicalIF":6.7,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13661805/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184359","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI and the future of teachers: predicting stakeholders' expectations of teacher job loss from AI-related institutional conditions.","authors":"Georgios P Georgiou","doi":"10.3389/frai.2026.1933015","DOIUrl":"https://doi.org/10.3389/frai.2026.1933015","url":null,"abstract":"<p><strong>Introduction: </strong>Artificial intelligence (AI) is transforming educational work, although claims that it will replace teachers often overlook the institutional conditions that make teacher job loss appear more or less likely. This study examined whether current AI-related institutional risk conditions predict stakeholder expectations of teacher job loss over the next decade.</p><p><strong>Methods: </strong>A purposive sample of 300 education and AI stakeholders rated the extent to which 11 conditions were currently present and provided a separate 0-10 score indicating the extent to which they expected AI to contribute to teacher job loss.</p><p><strong>Results: </strong>Overall job-loss expectations were low to moderate, and the institutional AI condition indicators showed good internal consistency. The machine learning model showed good predictive performance in the original held-out test set (<i>R</i> ² = 0.683) and stable performance across 100 repeated train-test splits (mean <i>R</i> ² = 0.598). The explainable AI analysis showed that the strongest contribution band consisted of automated feedback, large-scale AI adoption, and standardized content delivery.</p><p><strong>Discussion: </strong>The findings suggest that the extent to which stakeholders perceive AI-related institutional risk conditions to be currently present in educational settings is associated with the extent of teacher job loss they expect. They further indicate that this risk is not an inevitable technological outcome, but is embedded in a broader institutional ecology, with expectations clustering more strongly around direct forms of AI-mediated instructional restructuring than around indirect institutional or governance-related vulnerabilities.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1933015"},"PeriodicalIF":6.7,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13661954/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149196939","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SkinNet-X: a task coupled feature interaction framework for joint skin lesion segmentation and classification.","authors":"E Babu, S Murali","doi":"10.3389/frai.2026.1886221","DOIUrl":"10.3389/frai.2026.1886221","url":null,"abstract":"<p><p>Well trained segmentation and classification of skin lesions from dermoscopic images are crucial to an early detection of melanoma and other malignant skin disease. While many deep learning models reach impressive overall performance, most existing approaches treat segmentation and classification as separate tasks, leading to limited interaction between spatial lesion localization and semantic disease knowledge. This segregation results in a high variance prediction and would reduce the overall robustness of automated diagnostic systems. To mitigate this limitation, we proposed SkinNet-X: a unified multitask framework to reformulate the skin lesion segmentation and classification tasks jointly. We achieve this with a shared hierarchical feature representation to transfer knowledge between both tasks while introducing new Cross Task Consistency Learning (CTCL) strategy that enforces semantic consistency between lesion localization and disease prediction during joint optimization. In order to enhance the discriminative capability of feature expression and alleviate redundant representations, a dual stage multi label feature optimization module is further exploited, accompanied by an adaptive PolyLoss based optimization strategy stabilizing multitask learning. The effectiveness of the proposed SkinNet-X framework was evaluated using segmentation and classification experiments on the ISIC 2018, ISIC 2019, and HAM10000 benchmark datasets. We also performed additional analyses, namely ablation studies, computational complexity evaluation, sensitivity analysis and statistical validation to explore the contribution of each architectural component. Experimental results show that the proposed multitask learning strategy can consistently improve accuracy of lesion localization and performance of classification with robust generalizability across multiple datasets. These findings demonstrate that SkinNet-X is an effective and potentially useful framework for automated skin lesion analysis and computer-aided dermatological diagnosis.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1886221"},"PeriodicalIF":6.7,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13662072/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184666","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lung cancer classification and severity estimation using vision transformer-IPCNN with enhanced U-Net segmentation.","authors":"Nithya R, Vidhyapathi C M","doi":"10.3389/frai.2026.1942753","DOIUrl":"10.3389/frai.2026.1942753","url":null,"abstract":"<p><strong>Introduction: </strong>The death rate due to lung cancer is increasing because of late detection. Several automated lung cancer diagnostic techniques based on Machine Learning (ML) and Deep Learning (DL) have been developed for early detection. However, existing methods have limitations, including higher false-positive rates, lower sensitivity, and greater time consumption, which can affect early lung cancer diagnosis.</p><p><strong>Methods: </strong>To address these limitations, the Lung Cancer Classification and Severity Estimation-Improved Parallel Convolutional Neural Network-Vision Transformer (LCCSE-IPCNN-ViT) model is proposed. The framework consists of preprocessing, segmentation, classification, and severity estimation phases. First, the input images are preprocessed using the Improved Wiener Filtering (IWF) technique. The preprocessed images are then segmented using the Modified TransUNet (MTransUNet) model. Finally, the segmented images are classified and their severity is estimated using the IPCNN-ViT model, which combines the Improved Parallel Convolutional Neural Network (IPCNN) and Vision Transformer (ViT).</p><p><strong>Results: </strong>The proposed LCCSE-IPCNN-ViT model achieved an accuracy of 97.3%, sensitivity of 97.8%, specificity of 96.8%, precision of 96.8%, F-measure of 97.3%, and ROC-AUC of 98.5%.</p><p><strong>Discussion: </strong>The experimental results demonstrate that the proposed LCCSE-IPCNN-ViT model provides accurate and efficient lung cancer classification and severity estimation, indicating its potential for supporting early lung cancer diagnosis.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1942753"},"PeriodicalIF":6.7,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13661557/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184305","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prior-Enhanced Axial TransUNet: integrating MedSAM priors and sparse region-aware attention for kidney-tumor segmentation.","authors":"Siyuan Liang, Cheng Xu, Chao Lu, Jiewen Bao, Jiangchun Wu, Xin Gu, Bin Xu","doi":"10.3389/frai.2026.1848075","DOIUrl":"10.3389/frai.2026.1848075","url":null,"abstract":"<p><p>Accurate kidney and renal-tumor segmentation is challenging because lesion size, location, morphology, and boundary contrast vary substantially across abdominal CT scans. Most existing methods rely on a single form of local evidence and struggle to recover the boundaries of small lesions while maintaining global anatomical consistency. This paper proposes Prior-Enhanced Axial TransUNet (PEAT), which combines the broad medical representations of foundation models with task-specific discriminative learning. PEAT introduces three core components: Bottleneck Semantic Alignment (BSA) uses the frozen MedSAM image encoder as a training-stage feature reference and guides the semantic organization of the task network's bottleneck representation through an L1 alignment constraint; Dual-path Attention-enhanced Fusion (DAF) adaptively fuses CNN local features with MaxViT global contextual features by cascading coordinate attention and SENetV2; and Bi-granular Region-Aware Attention (BRA) achieves efficient non-local context modeling through Top-<i>k</i> window routing and patch-level sparse interaction. The MedSAM branch participates only during training, and only the task-specific network runs at inference. Under a setting in which all compared methods adopt the same preprocessing and evaluation protocol, PEAT achieves kidney and tumor Dice scores of 97.69% and 91.83% on KiTS21 and 96.80% and 92.58% on KiTS23, respectively. Across all four tasks, PEAT attains the highest Dice among the compared methods; ablation studies further show that the progressive integration of BSA, DAF, and BRA yields consistent gains, with the complete model outperforming all single and pairwise configurations. These results indicate that combining training-stage broad feature guidance with task-specific representation learning effectively improves kidney and renal-tumor segmentation and provides a viable route for the coordinated design of foundation-model priors and lightweight task networks.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1848075"},"PeriodicalIF":6.7,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13661798/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184391","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TL-CRANN: a transfer-learning-based hybrid CNN-RNN model with attention mechanism for cross-domain IoT attack detection.","authors":"L Sankar Ganesh, P Mohan Kumar","doi":"10.3389/frai.2026.1934604","DOIUrl":"https://doi.org/10.3389/frai.2026.1934604","url":null,"abstract":"<p><strong>Introduction: </strong>The rapid and pervasive growth of Internet of Things (IoT) devices has significantly expanded the attack surface of current network environments, making them vulnerable to advanced and evolving cyber threats. However, current intrusion detection systems (IDS) usually suffer from weak cross-domain generalization, high false positive rates, and poor performance against novel attacks.</p><p><strong>Methods: </strong>To overcome these issues, this paper proposes TL-CRANN, a transfer-learning-based hybrid deep learning framework utilizing a convolutional neural network (CNN), bidirectional long short-term memory (Bi-LSTM), and a multi-head self-attention mechanism for cross-domain IoT attack detection. Transfer learning is leveraged to transfer knowledge acquired from source-domain traffic and achieve rapid adaptation to heterogeneous target IoT environments with scarce labeled data. The proposed model is evaluated using three benchmark datasets: CICIDS2017, TON_IoT, and Bot-IoT.</p><p><strong>Results: </strong>Experimental results show that TL-CRANN achieved detection accuracies of 98.37%, 96.21%, and 97.14%, respectively, on the three datasets, and significantly outperformed existing state-of-the-art machine learning and deep learning baselines.</p><p><strong>Discussion: </strong>The results show that transfer learning improves the cross-domain detection performance and training cost. Moreover, the ablation analysis shows that CNN, Bi-LSTM and attention mechanisms-based modules are contribution to effectiveness of proposed framework which also confirms the applicability of TL-CRANN for cross-domain IoT attack detection.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1934604"},"PeriodicalIF":6.7,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13657973/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184747","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"From public innovation procurement requirements to AI-enabled socio-technical architecture: a longitudinal case study in industrialized housing.","authors":"Alberto Cerezo-Narváez, MCarmen Guerrero Delgado, Magdalena Ramírez-Peña, Daniel Sánchez-García, José Sánchez Ramos","doi":"10.3389/frai.2026.1962132","DOIUrl":"https://doi.org/10.3389/frai.2026.1962132","url":null,"abstract":"<p><p>Artificial intelligence (AI) is used in digital systems that combine data, computational methods, decision processes, and organizational arrangements. This paper examines how the requirements and institutional conditions of a public innovation procurement process are reflected in the development and specification of an AI-enabled socio-technical architecture for industrialized public housing delivery. The case concerns an AI-enabled system intended to support the design, assessment, and procurement preparation of industrialized housing within a public-sector delivery process. Using a longitudinal qualitative pre-deployment case-study design, we analyze a documentary corpus covering institutional procurement records, the initial technical proposal, its formal ex-ante assessment, and subsequent collaborative development. The architecture documented in the case comprises eight interconnected functional modules and three cross-cutting enablers: security, governance, and interoperability. It combines rule-based regulatory and contextual checking, generative and parametric design configuration, surrogate-model-supported performance estimation, cost and financial assessment, multi-objective optimization, multicriteria scenario comparison, and procurement-document generation. Across the corpus, we identify traceable correspondences between institutional requirements and functional, computational, governance, and interoperability arrangements, while keeping requirements, specifications, formal ex-ante assessment, subsequent development, and evidence requiring operational validation distinct. The contribution is an account of how AI-related and other computational capabilities are integrated with data, human decision points, governance arrangements, and procurement outputs within an evolving socio-technical architecture. We do not evaluate a completed system or establish operational effectiveness, scalability, or broader public outcomes. It provides a pre-deployment account of how procurement and institutional requirements correspond to the architecture of an AI-enabled decision-support system for industrialized housing delivery.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1962132"},"PeriodicalIF":6.7,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13657976/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Peircean framework for explainable and human-centred artificial intelligence.","authors":"Majeda Ahmad Omar, Ahmad Abdalla AlAhmad","doi":"10.3389/frai.2026.1962744","DOIUrl":"https://doi.org/10.3389/frai.2026.1962744","url":null,"abstract":"<p><p>This article develops a philosophical-conceptual framework for rethinking Explainable Artificial Intelligence (XAI) and Human-Centred Artificial Intelligence (HCAI) through the pragmatic semiotics of Charles Sanders Peirce. Rather than treating explainability solely as algorithmic transparency, the article conceptualizes it as a socially situated epistemic process through which AI-generated outputs are interpreted, evaluated, contested, and revised within human practices of inquiry. Drawing on Peirce's concepts of semiosis, abduction, inquiry, fallibilism, pragmaticism, and the community of inquiry, the framework distinguishes between computational processes performed by AI systems and the human and institutional practices through which their outputs acquire epistemic and practical significance. AI outputs may function as signs and may generate hypotheses that are functionally analogous to abductive possibilities; however, interpretation, evidential assessment, justification, ethical judgment, and responsibility remain situated within human communities of inquiry. The framework further translates Peircean concepts into practical principles for XAI, including contextual interpretation, uncertainty communication, contestability, human review, stakeholder participation, evidence-based revision, and continuous evaluation. A worked example involving generative AI illustrates how an AI-generated explanation can move through interpretation, questioning, evidence checking, contestation, and revision rather than being accepted as a final epistemic product. The article therefore positions Peircean pragmatism not as a replacement for technical XAI methods but as a philosophical and normative framework that complements them by clarifying the epistemological conditions of meaningful explanation, trustworthy human oversight, and responsible AI governance.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1962744"},"PeriodicalIF":6.7,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13658564/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184927","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Beyond text: quantifying the impact of figure-guided retrieval in biomedical RAG systems.","authors":"Raghav Kishore Venkatesh, G Sudhakaran, C U Om Kumar","doi":"10.3389/frai.2026.1861611","DOIUrl":"https://doi.org/10.3389/frai.2026.1861611","url":null,"abstract":"<p><strong>Introduction: </strong>Retrieval-Augmented Generation (RAG) has been established as a reliable paradigm for grounding large language model (LLM) outputs in factual evidence, acting as a safeguard in high-stakes domains such as biomedicine, where hallucination carries significant consequences. However, existing biomedical RAG systems rely predominantly on text-only corpora, overlooking the rich semantic information encoded in figures, charts, and other visual evidence. In this work, we investigate whether incorporating figure-based retrieval methods into a hybrid RAG pipeline meaningfully improves generation quality and retrieval effectiveness in the biomedical domain.</p><p><strong>Methods: </strong>We propose Figure-First Hybrid RAG (FFHRAG), a multi-stage retrieval architecture integrating multi-channel figure retrieval with hybrid text retrieval. Our architecture employs a <b>figure-first</b> discovery mechanism that retrieves relevant visual evidence, and utilizes it as a guiding signal for subsequent text selection. This strategy eliminates the need for complex embedding alignment, offering a <b>low-overhead</b> system that avoids <b>cross-modal embedding noise</b> at the final text chunk selection stage. The system is evaluated zero-shot on summary-type questions derived from BioASQ-13b, against a corpus spanning over 1.1 million PubMed Central text chunks and 82,000 figure-captions.</p><p><strong>Results: </strong>A progressive, multi-seed six-variant ablation study demonstrates that while the text-only baseline components establish a strong retrieval foundation-BM25 retriever yielding a Recall@10 of 0.7322 ± 0.0154 and the Cross-Encoder reranking securing an MRR of 0.8331 ± 0.0018, our proposed framework introduces isolated, targeted enhancements. Specifically, the localized PMCID-level boosting module refines retrieval ordering to an MRR of 0.8358 ± 0.0018, while prompt-level figure caption injection substantially enriches downstream generation quality, driving ROUGE-1 from 0.4068 ± 0.0037 to 0.4137 ± 0.0036 and ROUGE-L from 0.2623 ± 0.0018 to 0.2678 ± 0.0016 over the strong text-only baseline.</p><p><strong>Discussion: </strong>Our findings suggest that figures represent an underutilized retrieval modality in biomedical RAG systems, and that even lightweight figure-aware boosting mechanisms can produce measurable gains in answer quality without requiring multimodal fusion at the embedding level.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1861611"},"PeriodicalIF":6.7,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13658112/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence-driven personalization in digital marketing: effects on customer experience, purchase intention, and AI-supported managerial responsiveness.","authors":"Muchlis Abbas, Arnida Arnida, Muhammad Younus, Eliza Meiyani","doi":"10.3389/frai.2026.1962501","DOIUrl":"https://doi.org/10.3389/frai.2026.1962501","url":null,"abstract":"<p><strong>Introduction: </strong>Artificial intelligence (AI)-driven personalization has become a central strategy in digital marketing; however, existing studies provide a limited understanding of how AI simultaneously influences customer experience, consumer behavior, trust, and perceived organizational responsiveness.</p><p><strong>Objective: </strong>This study examines the effects of AI-driven personalization on customer experience, purchase intention, consumer trust, and perceived organizational responsiveness within the Indonesian digital market.</p><p><strong>Methodology: </strong>A quantitative approach was employed using survey data collected from 402 users of e-commerce and fintech platforms in Indonesia, which were analyzed using Structural Equation Modeling-Partial Least Squares (PLS-SEM).</p><p><strong>Results: </strong>The findings reveal that AI-driven personalization significantly enhances customer experience, which subsequently increases purchase intention. AI-based predictive analytics also demonstrates a significant positive relationship with perceived organizational responsiveness, indicating that consumers associate AI-enabled analytical capabilities with more timely and effective organizational responses to customer and market needs. Furthermore, customer experience serves as an important mediating mechanism, while perceived privacy risk weakens the positive relationship between AI-driven personalization and consumer trust.</p><p><strong>Conclusion: </strong>By aligning the organizational dimension with the perspective of consumer respondents, this study contributes an integrated understanding of how AI-driven personalization and predictive analytics shape consumer experiences and perceptions of organizational responsiveness in an emerging digital economy.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"9 ","pages":"1962501"},"PeriodicalIF":6.7,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13658568/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149184930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}