{"title":"A reliable approach for identifying acute lymphoblastic leukemia in microscopic imaging.","authors":"Mimosette Makem, Levente Tamas, Lucian Bușoniu","doi":"10.3389/frai.2025.1620252","DOIUrl":"https://doi.org/10.3389/frai.2025.1620252","url":null,"abstract":"<p><p>Leukemia is a deadly disease, and the patient's recovery rate is very dependent on early diagnosis. However, its diagnosis under the microscope is tedious and time-consuming. The advancement of deep convolutional neural networks (CNNs) in image classification has enabled new techniques in automated disease detection systems. These systems serve as valuable support and secondary opinion resources for laboratory technicians and hematologists when diagnosing leukemia through microscopic examination. In this study, we deployed a pre-trained CNN model (MobileNet) that has a small size and low complexity, making it suitable for mobile applications and embedded systems. We used the L1 regularization method and a novel dataset balancing approach, which incorporates HSV color transformation, saturation elimination, Gaussian noise addition, and several established augmentation techniques, to prevent model overfitting. The proposed model attained an accuracy of 95.33% and an F1 score of 0.95 when evaluated on the held-out test set extracted from the C_NMC_2019 public dataset. We also evaluated the proposed model by adding zero-mean Gaussian noise to the test images. The experimental results indicate that the proposed model is both efficient and robust, even when subjected to additional Gaussian noise. The comparison of the proposed MobileNet_M model's results with those of ALNet and various other existing models on the same dataset underscores its superior efficacy. The code is available for reproducing the experimental results at https://tamaslevente.github.io/ALLM/.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1620252"},"PeriodicalIF":4.7,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12310707/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144761600","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":"FLA-UNet: feature-location attention U-Net for foveal avascular zone segmentation in OCTA images.","authors":"Wei Li, Li Cao, He Deng","doi":"10.3389/frai.2025.1463233","DOIUrl":"https://doi.org/10.3389/frai.2025.1463233","url":null,"abstract":"<p><strong>Introduction: </strong>Since optical coherence tomography angiography (OCTA) is non-invasive and non-contact, it is widely used in the study of retinal disease detection. As a key indicator for retinal disease detection, accurate segmentation of foveal avascular zone (FAZ) has an important impact on clinical application. Although the U-Net and its existing improvement methods have achieved good performance on FAZ segmentation, their generalization ability and segmentation accuracy can be further improved by exploring more effective improvement strategies.</p><p><strong>Methods: </strong>We propose a novel improved method named Feature-location Attention U-Net (FLA-UNet) by introducing new designed feature-location attention blocks (FLABs) into U-Net and using a joint loss function. The FLAB consists of feature-aware blocks and location-aware blocks in parallel, and is embed into each decoder of U-Net to integrate more marginal information of FAZ and strengthen the connection between target region and boundary information. The joint loss function is composed of the cross-entropy loss (CE loss) function and the Dice coefficient loss (Dice loss) function, and by adjusting the weights of them, the performance of the network on boundary and internal segmentation can be comprehensively considered to improve its accuracy and robustness for FAZ segmentation.</p><p><strong>Results: </strong>The qualitative and quantitative comparative experiments on the three datasets of OCTAGON, FAZID and OCTA-500 show that, our proposed FLA-UNet achieves better segmentation quality, and is superior to other existing state-of-the-art methods in terms of the MIoU, ACC and Dice coefficient.</p><p><strong>Discussion: </strong>The proposed FLA-UNet can effectively improve the accuracy and robustness of FAZ segmentation in OCTA images by introducing feature-location attention blocks into U-Net and using a joint loss function. This has laid a solid theoretical foundation for its application in auxiliary diagnosis of fundus diseases.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1463233"},"PeriodicalIF":4.7,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12310713/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144761601","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}
David Antony Selby, Rashika Jakhmola, Maximilian Sprang, Gerrit Großmann, Hind Raki, Niloofar Maani, Daria Pavliuk, Jan Ewald, Sebastian Vollmer
{"title":"Visible neural networks for multi-omics integration: a critical review.","authors":"David Antony Selby, Rashika Jakhmola, Maximilian Sprang, Gerrit Großmann, Hind Raki, Niloofar Maani, Daria Pavliuk, Jan Ewald, Sebastian Vollmer","doi":"10.3389/frai.2025.1595291","DOIUrl":"https://doi.org/10.3389/frai.2025.1595291","url":null,"abstract":"<p><strong>Background: </strong>Biomarker discovery and drug response prediction are central to personalized medicine, driving demand for predictive models that also offer biological insights. Biologically informed neural networks (BINNs), also referred to as visible neural networks (VNNs), have recently emerged as a solution to this goal. BINNs or VNNs are neural networks whose inter-layer connections are constrained based on prior knowledge from gene ontologies and pathway databases. These sparse models enhance interpretability by embedding prior knowledge into their architecture, ideally reducing the space of learnable functions to those that are biologically meaningful.</p><p><strong>Methods: </strong>This systematic review-the first of its kind-identified 86 recent papers implementing BINNs/VNNs. We analyzed these papers to highlight key trends in architectural design, data sources and evaluation methodologies.</p><p><strong>Results: </strong>Our analysis reveals a growing adoption of BINNs/VNNs. However, this growth is apparently juxtaposed with a lack of standardized, terminology, computational tools and benchmarks.</p><p><strong>Conclusion: </strong>BINNs/VNNs represent a promising approach for integrating biological knowledge into predictive models for personalized medicine. Addressing the current deficiencies in standardization and tooling is important for widespread adoption and further progress in the field.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1595291"},"PeriodicalIF":4.7,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12310660/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144761602","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}
Heily Consepción Portocarrero Ramos, Omer Cruz Caro, Einstein Sánchez Bardales, Lenin Quiñones Huatangari, Jonathan Alberto Campos Trigoso, Jorge Luis Maicelo Guevara, River Chávez Santos
{"title":"Artificial intelligence skills and their impact on the employability of university graduates.","authors":"Heily Consepción Portocarrero Ramos, Omer Cruz Caro, Einstein Sánchez Bardales, Lenin Quiñones Huatangari, Jonathan Alberto Campos Trigoso, Jorge Luis Maicelo Guevara, River Chávez Santos","doi":"10.3389/frai.2025.1629320","DOIUrl":"https://doi.org/10.3389/frai.2025.1629320","url":null,"abstract":"<p><p>Artificial intelligence (AI) has emerged as a transformative technology in multiple areas, including the labor market. Its incorporation into organizations redefines professional profiles, required skills, and employability conditions. In this context, it is essential to understand how university graduates are preparing to face these changes and what role their AI skills play in their integration into the workforce. The study aimed to analyze the level of AI skills and their impact on the employability of university graduates through a quantitative and descriptive design. A survey was conducted with a sample of 148 undergraduate and graduate graduates. The data were analyzed using descriptive statistics and visualized using graphs. The results indicated that graduates who report greater knowledge and more frequent use of AI tools, especially generative ones such as ChatGPT, are more likely to be employed in areas related to their majors and to perceive higher productivity and better professional alignment. However, a generational gap in digital skills was also identified, as well as a widespread feeling of insufficient preparation for the challenges of the current labor market. The conclusion is that AI skills are consolidating as a key differentiating factor in employability and that their formal incorporation into university curricula is urgently needed. The implications of the study point to the need for an educational transformation that integrates AI as a transversal skill, promotes ongoing teacher training, and fosters policies that guarantee inclusive education aligned with the challenges of the digital age.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1629320"},"PeriodicalIF":4.7,"publicationDate":"2025-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12307348/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144754648","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":"Ethical theories, governance models, and strategic frameworks for responsible AI adoption and organizational success.","authors":"Mitra Madanchian, Hamed Taherdoost","doi":"10.3389/frai.2025.1619029","DOIUrl":"https://doi.org/10.3389/frai.2025.1619029","url":null,"abstract":"<p><p>As artificial intelligence (AI) becomes integral to organizational transformation, ethical adoption has emerged as a strategic concern. This paper reviews ethical theories, governance models, and implementation strategies that enable responsible AI integration in business contexts. It explores how ethical theories such as utilitarianism, deontology, and virtue ethics inform practical models for AI deployment. Furthermore, the paper investigates governance structures and stakeholder roles in shaping accountability and transparency, and examines frameworks that guide strategic risk assessment and decision-making. Emphasizing real-world applicability, the study offers an integrated approach that aligns ethics with performance outcomes, contributing to organizational success. This synthesis aims to support firms in embedding responsible AI principles into innovation strategies that balance compliance, trust, and value creation.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1619029"},"PeriodicalIF":4.7,"publicationDate":"2025-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12307427/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144754649","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}
V L Sowmya, A Bharathi Malakreddy, Santhi Natarajan, N Prathik, I S Rajesh
{"title":"Spectral Entropic Radiomics Feature Extraction (SERFE): an adaptive approach for glioblastoma disease classification.","authors":"V L Sowmya, A Bharathi Malakreddy, Santhi Natarajan, N Prathik, I S Rajesh","doi":"10.3389/frai.2025.1583079","DOIUrl":"https://doi.org/10.3389/frai.2025.1583079","url":null,"abstract":"<p><strong>Introduction: </strong>Radiomics-based glioblastoma classification demands feature extraction techniques that can effectively capture tumor heterogeneity while maintaining computational efficiency. Conventional tools such as PyRadiomics and CaPTk rely on extensive handcrafted feature sets, which often result in redundancy and necessitate further optimization steps.</p><p><strong>Methods: </strong>This study proposes a novel framework, Spectral Entropic Radiomics Feature Extraction (SERFE), which integrates spectral frequency decomposition, entropy-driven feature selection, and graph-based spatial encoding. SERFE decomposes voxel intensity fluctuations into spectral signatures, employs entropy-based weighting to prioritize informative features, and preserves spatial topology through graph-based modeling. The method was evaluated using the public TCIA glioblastoma dataset.</p><p><strong>Results: </strong>SERFE generated a refined feature set of 350 radiomic features from an initial pool of 2,260, achieving a 92% stability score and 91.7% classification accuracy. This performance surpasses traditional radiomics methods in both predictive accuracy and feature compactness.</p><p><strong>Discussion: </strong>The results demonstrate SERFE's capacity to enhance tumor characterization and streamline radiomics pipelines without requiring post-extraction feature reduction. Its compatibility with existing clinical workflows makes it a promising tool for future neuro-oncology applications.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1583079"},"PeriodicalIF":4.7,"publicationDate":"2025-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12307471/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144754596","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}
Shikha Chaganti, Vivek Singh, Alasdair Edward Gent, Rishikesan Kamaleswaran, Ali Kamen
{"title":"Evaluating the impact of common clinical confounders on performance of deep-learning-based sepsis risk assessment.","authors":"Shikha Chaganti, Vivek Singh, Alasdair Edward Gent, Rishikesan Kamaleswaran, Ali Kamen","doi":"10.3389/frai.2025.1452471","DOIUrl":"10.3389/frai.2025.1452471","url":null,"abstract":"<p><strong>Introduction: </strong>Early identification of sepsis in the emergency department using machine learning remains a challenging problem, primarily due to the lack of a gold standard for sepsis diagnosis, the heterogeneity in clinical presentations, and the impact of confounding conditions.</p><p><strong>Methods: </strong>In this work, we present a deep-learning-based predictive model designed to enable early detection of patients at risk of developing sepsis, using data from the first 24 h of admission. The model is based on routine blood test results commonly performed on patients, including CBC (Complete Blood Count), CMP (Comprehensive Metabolic Panel), lipid panels, vital signs, age, and sex. To address the challenge of label uncertainty as a part of the training process, we explore two different definitions, namely, Sepsis-3 and Adult Sepsis Event. We analyze the advantages and limitations of each in the context of patient clinical parameters and comorbidities. We specifically examine how the quality of the ground truth label influences the performance of the deep learning system and evaluate the effect of a consensus-based approach that incorporates both definitions. We also evaluated the model's performance across sub-cohorts, including patients with confounding comorbidities (such as chronic kidney, liver disease, and coagulation disorders) and those with infections confirmed by billing codes.</p><p><strong>Results: </strong>Our results show that the consensus-based model identifies at-risk patients in the first 24 h with 83.7% sensitivity, 80% specificity, 36% PPV, 97% NPV, and an AUC of 0.9. Our cohort-wise analysis revealed a high PPV (77%) in infection-confirmed subgroups and a drop in specificity across cohorts with confounding comorbidities (47-70%).</p><p><strong>Discussion: </strong>This work highlights the limitations of retrospective sepsis definitions and underscores the need for tailored approaches in automated sepsis detection, particularly when dealing with patients with confounding comorbidities.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1452471"},"PeriodicalIF":4.7,"publicationDate":"2025-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12305701/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144745278","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":"CMDMamba: dual-layer Mamba architecture with dual convolutional feed-forward networks for efficient financial time series forecasting.","authors":"Zhenkai Qin, Baozhong Wei, Yujia Zhai, Ziqian Lin, Xiaochuan Yu, Jingxuan Jiang","doi":"10.3389/frai.2025.1599799","DOIUrl":"10.3389/frai.2025.1599799","url":null,"abstract":"<p><strong>Introduction: </strong>Transformer models have demonstrated remarkable performance in financial time series forecasting. However, they suffer from inefficiencies in computational efficiency, high operational costs, and limitations in capturing temporal dependencies.</p><p><strong>Methods: </strong>To address these challenges, we propose the CMDMamba model, which is based on the Mamba architecture of state-space models (SSMs) and achieves near-linear time complexity. This significantly enhances the real-time data processing capability and reduces the deployment costs for risk management systems. The CMDMamba model employs a dual-layer Mamba structure that effectively captures price fluctuations at both the micro- and macrolevels in financial markets and integrates an innovative Dual Convolutional Feedforward Network (DconvFFN) module. This module is able to effectively capture the correlations between multiple variables in financial markets. By doing so, it provides more accurate time series modeling, optimizes algorithmic trading strategies, and facilitates investment portfolio risk warnings.</p><p><strong>Results: </strong>Experiments conducted on four real-world financial datasets demonstrate that CMDMamba achieves a 10.4% improvement in prediction accuracy for multivariate forecasting tasks compared to state-of-the-art models.</p><p><strong>Discussion: </strong>Moreover, CMDMamba excels in both predictive accuracy and computational efficiency, setting a new benchmark in the field of financial time series forecasting.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1599799"},"PeriodicalIF":4.7,"publicationDate":"2025-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12303894/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144745277","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}
Yuriy Vasilev, Anastasia Pamova, Tatiana Bobrovskaya, Anton Vladzimirskyy, Olga Omelyanskaya, Elena Astapenko, Artem Kruchinkin, Novik Vladimir, Kirill Arzamasov
{"title":"Outliers and anomalies in training and testing datasets for AI-powered morphometry-evidence from CT scans of the spleen.","authors":"Yuriy Vasilev, Anastasia Pamova, Tatiana Bobrovskaya, Anton Vladzimirskyy, Olga Omelyanskaya, Elena Astapenko, Artem Kruchinkin, Novik Vladimir, Kirill Arzamasov","doi":"10.3389/frai.2025.1607348","DOIUrl":"10.3389/frai.2025.1607348","url":null,"abstract":"<p><strong>Introduction: </strong>Creating training and testing datasets for machine learning algorithms to measure linear dimensions of organs is a tedious task. There are no universally accepted methods for evaluating outliers or anomalies in such datasets. This can cause errors in machine learning and compromise the quality of end products. The goal of this study is to identify optimal methods for detecting organ anomalies and outliers in medical datasets designed to train and test neural networks in morphometrics.</p><p><strong>Methods: </strong>A dataset was created containing linear measurements of the spleen obtained from CT scans. Labelling was performed by three radiologists. The total number of studies included in the sample was <i>N</i> = 197 patients. Using visual methods (1.5 interquartile range; heat map; boxplot; histogram; scatter plot), machine learning algorithms (Isolation forest; Density-Based Spatial Clustering of Applications with Noise; K-nearest neighbors algorithm; Local outlier factor; One-class support vector machines; EllipticEnvelope; Autoencoders), and mathematical statistics (z-score, Grubb's test; Rosner's test).</p><p><strong>Results: </strong>We identified measurement errors, input errors, abnormal size values and non-standard shapes of the organ (sickle-shaped, round, triangular, additional lobules). The most effective methods included visual techniques (including boxplots and histograms) and machine learning algorithms such is OSVM, KNN and autoencoders. A total of 32 outlier anomalies were found.</p><p><strong>Discussion: </strong>Curation of complex morphometric datasets must involve thorough mathematical and clinical analyses. Relying solely on mathematical statistics or machine learning methods appears inadequate.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1607348"},"PeriodicalIF":4.7,"publicationDate":"2025-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12303909/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144745279","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}
Dalia Alammari, Elaf Alamari, Rawan Alamri, Raneem Alharbi, Jumana Felimban, Jana Aljohani
{"title":"Assessing medical students' attitudes, performance, and usage of ChatGPT in Jeddah, Saudi Arabia.","authors":"Dalia Alammari, Elaf Alamari, Rawan Alamri, Raneem Alharbi, Jumana Felimban, Jana Aljohani","doi":"10.3389/frai.2025.1577911","DOIUrl":"10.3389/frai.2025.1577911","url":null,"abstract":"<p><strong>Background: </strong>ChatGPT, an advanced AI language model, has the potential to significantly enhance medical education by supporting clinical decision-making, facilitating knowledge acquisition, and improving learning outcomes. However, there remains a gap in understanding the risks and concerns surrounding the use of ChatGPT, as well as its impact on the quality of medical education in Jeddah, Saudi Arabia. This study investigates the role of ChatGPT in local medical education, offering insights into medical students' attitudes and performance concerning the integration of AI into the medical curriculum.</p><p><strong>Methods: </strong>This cross-sectional study was conducted from March to May 2024. It was approved by the institutional review board. An online survey was distributed to medical students in Jeddah to collect data on their awareness of ChatGPT and its impact on their attitudes and performance. The survey, which included 28 items across 2 sections was developed. For data analysis, Statistical Package for Social Sciences (SPSS) software (version 27.0) was used to conduct statistical analysis.</p><p><strong>Results: </strong>The final sample comprised 420 participants, of whom 84.3% had heard of ChatGPT, while 74.9% had used it prior to the study. The majority of participants were aged 18-21 (50.5%). Higher GPA and academic progression were significantly associated with greater awareness and performance related to ChatGPT. Additionally, privacy concerns, willingness to incorporate ChatGPT into learning and research, and perceptions of its ease of use were significantly correlated, with a <i>p</i>-value < 0.05.</p><p><strong>Conclusion: </strong>The findings indicate a generally positive perception of ChatGPT among medical students, particularly as they progress in their studies. Associations were observed between ChatGPT usage and students' academic standing and attitudes toward AI. While the ease of use was appreciated, concerns regarding privacy, ethical implications, and data security were also prominent, reflecting global trends. Further longitudinal and experimental research is necessary to better understand the educational implications of ChatGPT and to ensure its responsible integration into medical curricula.</p>","PeriodicalId":33315,"journal":{"name":"Frontiers in Artificial Intelligence","volume":"8 ","pages":"1577911"},"PeriodicalIF":4.7,"publicationDate":"2025-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12302260/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144733561","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}