Balancing Privacy and Progress: A Review of Privacy Challenges, Systemic Oversight, and Patient Perceptions in AI-Driven Healthcare

Applied Sciences, Год журнала: 2024, Номер 14(2), С. 675 - 675

Опубликована: Янв. 12, 2024

Integrating Artificial Intelligence (AI) in healthcare represents a transformative shift with substantial potential for enhancing patient care. This paper critically examines this integration, confronting significant ethical, legal, and technological challenges, particularly privacy, decision-making autonomy, data integrity. A structured exploration of these issues focuses on Differential Privacy as critical method preserving confidentiality AI-driven systems. We analyze the balance between privacy preservation practical utility data, emphasizing effectiveness encryption, Privacy, mixed-model approaches. The navigates complex ethical legal frameworks essential AI integration healthcare. comprehensively examine rights nuances informed consent, along challenges harmonizing advanced technologies like blockchain General Data Protection Regulation (GDPR). issue algorithmic bias is also explored, underscoring urgent need effective detection mitigation strategies to build trust. evolving roles decentralized sharing, regulatory frameworks, agency are discussed depth. Advocating an interdisciplinary, multi-stakeholder approach responsive governance, aims align principles, prioritize patient-centered outcomes, steer towards responsible equitable enhancements

Язык: Английский

---

Cybersecurity for Sustainable Smart Healthcare: State of the Art, Taxonomy, Mechanisms, and Essential Roles

Deleted Journal, Год журнала: 2024, Номер 4(2), С. 20 - 62

Опубликована: Май 23, 2024

Cutting-edge technologies have been widely employed in healthcare delivery, resulting transformative advances and promising enhanced patient care, operational efficiency, resource usage. However, the proliferation of networked devices data-driven systems has created new cybersecurity threats that jeopardize integrity, confidentiality, availability critical data. This review paper offers a comprehensive evaluation current state context smart healthcare, presenting structured taxonomy its existing cyber threats, mechanisms essential roles. study explored (SHSs). It identified discussed most pressing attacks SHSs face, including fake base stations, medjacking, Sybil attacks. examined security measures deployed to combat SHSs. These include cryptographic-based techniques, digital watermarking, steganography, many others. Patient data protection, prevention breaches, maintenance SHS integrity are some roles ensuring sustainable healthcare. The long-term viability depends on constant assessment risks harm providers, patients, professionals. aims inform policymakers, practitioners, technology stakeholders about imperatives best practices for fostering secure resilient ecosystem by synthesizing insights from multidisciplinary perspectives, such as cybersecurity, management, sustainability research. Understanding recent is controlling escalating networks encouraging intelligent delivery.

Язык: Английский

---

A Fog-Based Privacy-Preserving Federated Learning System for Smart Healthcare Applications

Electronics, Год журнала: 2023, Номер 12(19), С. 4074 - 4074

Опубликована: Сен. 28, 2023

During the COVID-19 pandemic, urgency of effective testing strategies had never been more apparent. The fusion Artificial Intelligence (AI) and Machine Learning (ML) models, particularly within medical imaging (e.g., chest X-rays), holds promise in smart healthcare systems. Deep (DL), a subset AI, has exhibited prowess enhancing classification accuracy, crucial aspect expediting diagnosis. However, journey to harness DL’s potential is rife with challenges: notably, intricate landscape data privacy. Striking balance between utilizing patient for insights while upholding privacy formidable. Federated (FL) emerges as solution by enabling collaborative model training across decentralized sources, thus bypassing centralization preserving This study presents tailored, FL architecture screening via X-ray images. Designed facilitate cooperation among institutions, framework ensures remain localized, eliminating need direct sharing. Addressing imbalanced non-identically distributed data, robust solution. Implementation entails localized fog-computing-based models. Localized models utilize Convolutional Neural Networks (CNNs) on institution-specific datasets, model, refined iteratively, takes precedence final classification. Intriguingly, global fortified fog computing, frontrunner after weight refinement, surpassing local Validation COLAB platform gauges model’s performance through metrics such precision, recall, F1-score. Remarkably, proposed excels these metrics, solidifying its efficacy. research navigates confluence FL, imaging, unveiling that could reshape delivery. enriches scientific discourse addressing learning carries implications enhanced care.

Язык: Английский

---

Privacy-Preserving Federated Learning for Intrusion Detection in IoT Environments: A Survey

IEEE Access, Год журнала: 2024, Номер 12, С. 127018 - 127050

Опубликована: Янв. 1, 2024

Язык: Английский

---

Big Data’s Impact on Healthcare and Bioinformatics

Studies in big data, Год журнала: 2025, Номер unknown, С. 23 - 51

Опубликована: Янв. 1, 2025

Язык: Английский

---

Challenges for the Integration of Artificial Intelligence in Healthcare Services: A Decision-Making Approach

OPUS Toplum Araştırmaları Dergisi, Год журнала: 2025, Номер 22(1), С. 23 - 32

Опубликована: Фев. 16, 2025

This study aims to elucidate the interdependent effects of challenges and risks using artificial intelligence in healthcare sector. The ten obtained by literature were assessed five professionals involved managing health. Participants selected based on having at least years academic or professional experience participants made their judgments topic structured forms. DEMATEL (The Decision-Making Trial Evaluation Laboratory) technique investigated cause-effect relationships between identified integration challenges. According analysis results terms degree importance, safety security risk (SSR) is ranked first place, inadequate patient assessments (IPRA), data quality (DQR), verifiability (VR), stakeholders perceived mistrust (SPM), (IC), ethical considerations (EC), algorithm/decision-making bias (AMB) job displacement (JDR) are following places. In addition, DQR, AMB, SSR, VR, IPRA, DPR causal variables; EC, IC, JDR, SPM regarded as effects. These factors highlight need for robust mechanisms ensure integrity data, accuracy assessments, transparency decision-making processes AI. Negative impacts ethics, inclusion, employment, trust will likely be reduced addressing root causes, such quality, assessment, algorithmic bias, developing policies address them.

Язык: Английский

---

Blockchain-Based Decentralized Identity Systems: A Survey of Security, Privacy, and Interoperability

Опубликована: Март 27, 2025

Blockchain technology, a decentralized and immutable ledger, has transformed identity access management (IAM) by enhancing security, privacy, trust in digital ecosystems. Ensuring safe authentication data integrity is made possible its integration with sophisticated cryptographic techniques like zero-knowledge proofs (ZKPs) public- key infrastructure (PKI). Other methods include verifiable credentials (VCs) identifiers (DIDs). This paper provides comprehensive analysis of blockchain-based IAM systems, comparing leading blockchain platforms, including Ethereum, Hyperledger Indy, IOTA, IoTeX, management. The role mitigating identity-related threats, such as theft unauthorized access, explored through decentralization, immutability, smart contract automation. Additionally, security enhancements, mechanisms that strengthen solutions privacy-preserving authentication, are examined. potential to establish self-sovereign framework fosters trust, scalability, ecosystems highlighted, paving the way for next generation solutions.

Язык: Английский

---

Privacy-Preserving AI for Medical Applications

IGI Global eBooks, Год журнала: 2025, Номер unknown, С. 125 - 148

Опубликована: Апрель 24, 2025

The sudden infusion of artificial intelligence (AI) into the medical sector requires strong frameworks that address regulatory requirements, ethical issues, privacy and concerns related to interoperability. current research conceptualizes a dynamic system AI governance changes dynamically along with advancements in AI, avoiding constraints fixed models governance. In contrast conventional methods, model incorporates adaptive compliance processes, explainable models, user-managed data-sharing systems enable transparency trustworthiness. It also uses federated learning methods secure scalable adoption healthcare avoid data heterogeneity challenges. includes enhanced protection less homomorphic encryption optimally utilized blockchain, reducing computation overhead allowing practical application.

Язык: Английский

---

FED-LIFE: Ghost LinkNet Enabled Federated Learning for Anomaly Detection in Smart Intensive Care Unit based on IOMT

Опубликована: Май 13, 2025

Intensive Care Unit (ICU) patient monitoring plays a vital role in ensuring the safety and well-being of critically ill patients by providing continuous real-time insights into their health status. The integration Internet Medical Things (IoMT) devices ICU including wearable sensors remote tools, enables seamless collection transmission data, allowing for tracking signs. Federated learning (FL) enhances this process utilizing decentralized data to improve model generalization while maintaining privacy. However, FL-based faced challenges high delays decision-making due centralized processing, significant execution time caused need transfer large volumes data. This research proposes novel FEDerated learning-based LIFE saving system (FED-LIFE) effective timely services patients. FED-LIFE initially trains local models Ghostnet combined Enhanced LinkNet (Ghost_EliNet) which combines GhostNet LinkNet, tuning Ghost_EliNet Red Deer Optimization (RDO) algorithm is employed accurate service allocation. suggested approach implemented Python programming. efficacy developed evaluated several metrics namely Precision, recall, f1-score, accuracy, delay, throughput, time. proposed method achieves lowest delay 22 seconds 50 Whereas existing FEDSDM, Deep-CFL, FL-IRL attain 45 seconds, 37 35 respectively.

Язык: Английский

---

Secure Federated GenAI for Healthcare IoT Networks

International Journal of Latest Technology in Engineering Management & Applied Science, Год журнала: 2025, Номер 14(4), С. 785 - 794

Опубликована: Май 16, 2025

Abstract: The combination of Healthcare Internet Things (HIoT) and Generative Artificial Intelligence (GenAI) has offered an eminent opportunity to develop highly personalized, real-time healthcare offerings. However, the need for data security, cloud dependency, sensitivity in handling medical information have come way mass adoption. This paper proposes a secure Federated GenAI framework HIoT networks allow distributed training generation AI models directly on edge devices like wearables mobile health sensors. By combining federated learning protocols with privacy-preserving mechanisms generative AI, system minimizes transmitting raw centralized servers. architecture asserts sovereignty, risk breaches, preserves performance operating across nodes. experimental evaluation showed achieves competitive accuracy levels monitoring tasks along strong privacy guarantees communication efficiency. Therefore, our results signal that can be credible base developing scalable ethical AI-driven systems, particularly settings are resource-constrained or laden regulations.

Язык: Английский

---