Privacy Risks Research Articles

Exploring the Impact of Social Robot Design Characteristics on Users’ Privacy Concerns: Evidence from PLS-SEM and FsQCA

Although an increasing number of studies explore the factors influencing users’ privacy concerns regarding social robots, the existing understanding of this issue remains largely fragmented. Previous studies have mainly focused on the "net effect" between variables, leaving the complexity of causal configurations, and the holistic impact of design characteristics of social robots on user privacy concerns remains unclear. Based on the Stimuli-Organism-Response (S-O-R) framework and Communication Privacy Management Theory (CPMT), this study integrates social robot design characteristics such as Anthropomorphism, Warmth, Competence, and Transparency into causal configurations, and uses Perceived Privacy Risk and Perceived Privacy Control as mediating variables to propose a Comprehensive conceptual model. Based on valid data from a sample of 198 Chinese social robot users, this study conducted empirical analyses of the conceptual model using Partial Least Squares Structural Equation Modeling (PLS-SEM) and Fuzzy-set Qualitative Comparative Analysis (FsQCA). PLS-SEM results show that anthropomorphism, warmth, competence, and transparency are key factors influencing privacy concerns, and perceived privacy risk mediates the relationship between warmth, information transparency, and privacy concerns. The FsQCA results further validated the findings of PLS-SEM and identified five configurations of factor combinations that led to higher levels of user privacy concerns. Among them, the combination of high anthropomorphism design, high competence, and low warmth of social robots is the core configuration that leads to users’ privacy concerns. Overall, this study broadens our understanding of social robot users’ privacy concerns and reveals the causal complexity behind social robot users’ privacy concerns. It provides some theoretical and practical insights for subsequent scholars and designers.

Improving Athlete Data Protection: Tackling Privacy and Economic Risks in Digital Security

Improving Athlete Data Protection: Tackling Privacy and Economic Risks in Digital Security

PDAC-SL: A PUF-enabled drone access control technique for smart logistics

The conventional logistics’ paradigm is being replaced with smarter ways of supply chain management to readily serve the customers with quick modes of delivery in affordable operational costs. The weak infrastructure in rural terrain and congested roads in urban cities act as a big impediment in the successful accomplishment of last mile delivery down the lane across the complete chain of logistics. The traditional logistics has been evolved and improved in many respects, the last mile delivery is still plagued by security and privacy risks in logistics. The use of drones for package deliveries may offer promising gains in terms of efficiency. Nonetheless, drones undergo security and privacy risks besides physical threats for being deployed in public places communicating on open channels. Thus, more efforts are needed to secure the communication of inducted drones in delivery assignments. We propose a novel PUF-enabled drone access control mechanism PDAC-SL for smart logistics. By the careful employment of symmetric crypto-primitives, the contributed protocol not only supports low end devices but stands resistant to physical drone and other known threats besides preserving anonymity and untraceability features. The RoR model-based formal security analysis and performance analysis demonstrate distinct security properties and applicability of the PDAC-SL.

Consumer acceptance of social robots in domestic settings: A human-robot interaction perspective

Consumer acceptance of social robots in retail and service settings is an emerging field in marketing research. Social Home Robots (e.g., Jibo, Cozmo, and Kuri) serving various groups in domestic settings show significant retail and consumer service potential. However, most current research focuses on social robots that serve older or disabled individuals in domestic settings, leaving a gap in understanding consumer acceptance of Social Home Robots, particularly across different cultures. From the Human-Robot Interaction perspective, this article examines consumers' intentions to use Social Home Robots through a sequential mixed-method approach. Our findings indicate that consumers' intentions to use these robots are primarily influenced by key factors in Human-Robot Interaction: belief in social robot benefits, concern about privacy risks, trust, the robots' overall anthropomorphic design, and consumers' self-construal tendency. Furthermore, trust mediates the relationships between belief in social support, privacy concerns, independent self-construal, and usage intention. This study explores consumer acceptance of social robots serving diverse groups rather than special groups in domestic settings, contributing to the literature on social/service robots. It also offers business guidelines for designing and marketing Social Home Robots in retail.

Emerging market dynamics: risk perceptions, perceived usefulness and E-pharmacy adoption

Purpose Despite the digital penetration in the larger consumer market, the adoption of e-pharmacy services is reported to be very low in emerging economies like India. In this context, the purpose of the current study is to examine the influence of consumers’ risk perception on their intention to purchase medicines/pharmaceuticals through online pharmacies. Furthermore, the study seeks to understand to what extent the perceived usefulness of online pharmacy mediates the relationship between different dimensions of perceived risk and purchase intention. Design/methodology/approach The study was conducted among a sample of 800 consumers in India who are familiar with online e-commerce. To ensure the homogeneity of the sample and hence the generalizations of results, inclusion criteria were set as not to include respondents who have made a prior purchase through e-pharmacy services. Self-reporting questionnaires were administered among the respondents, who were selected through a purposive and convenience sampling method. Measurement modeling and path analysis were done using IBM SPSS 23.0 and AMOS 24.0 to test the hypotheses and draw inferences. Findings Results revealed that consumers’ risk perceptions, such as financial, product, source and privacy risks, had a significant direct and indirect effect on their intention to purchase medicines through online pharmacies. India is the major universal provider of generic medicines. The insights gained from this study can help policymakers, corporates, consumers, distributors, retailers and marketing managers to frame effective strategies for improved usage of online platforms for procuring medications. Originality/value This study is pioneering in conceptualizing and testing a theoretical model linking consumers’ risk perceptions, perceived usefulness and intention to purchase through online pharmacies, particularly in the context of an emerging economy like India. Implications regarding facilitating and nurturing a conducive platform for online purchasing medicines and its outcomes are elaborated, thus striving to fill a gap in the existing literature. By examining the proposed framework through the lens of the technology acceptance model and theory of risk perception, this study seeks to add to the emergent literature on online pharmacies, especially in emerging economies with huge market potential.

Hierarchical Progressive Image Forgery Detection and Localization Method Based on UNet

The rapid development of generative technologies has made the production of forged products easier, and AI-generated forged images are increasingly difficult to accurately detect, posing serious privacy risks and cognitive obstacles to individuals and society. Therefore, constructing an effective method that can accurately detect and locate forged regions has become an important task. This paper proposes a hierarchical and progressive forged image detection and localization method called HPUNet. This method assigns more reasonable hierarchical multi-level labels to the dataset as supervisory information at different levels, following cognitive laws. Secondly, multiple types of features are extracted from AI-generated images for detection and localization, and the detection and localization results are combined to enhance the task-relevant features. Subsequently, HPUNet expands the obtained image features into four different resolutions and performs detection and localization at different levels in a coarse-to-fine cognitive order. To address the limited feature field of view caused by inconsistent forgery sizes, we employ three sets of densely cross-connected hierarchical networks for sufficient interaction between feature images at different resolutions. Finally, a UNet network with a soft-threshold-constrained feature enhancement module is used to achieve detection and localization at different scales, and the reliance on a progressive mechanism establishes relationships between different branches. We use ACC and F1 as evaluation metrics, and extensive experiments on our method and the baseline methods demonstrate the effectiveness of our approach.

Development of a Enhanced System Review for Public Agency Multi-Sectoral Integrated Data Systems

While traditional security assessments (such as mandated external security reviews which include a penetration test tied to a security framework) are a critical component to securing your data infrastructure, the complexities and governance structures around integrated data systems that involve multiple public sector agencies require a more comprehensive framework to address comprehensive risk, regulatory, and governance considerations that are critical to the confidentiality of the data within these systems. In January of 2024, the Data Integration Support Center (DISC) created and piloted an enhanced system review that includes components that are often overlooked or oversimplified when working with today's technologies such as cloud-based secure enclaves. By following up on a mandated security review, an enhanced system review can help a maturing Integrated Data System (IDS) address both security and privacy risks as it moves forward. An enhanced system review looks primarily, though not necessarily exclusively, at the following aspects of an IDS, principally in the interest of ensuring privacy: IDS alignment with its legal framework and other structuring documents (e.g., policies, procedures, interagency agreements) Training to support privacy and security Use of technology to prevent or mitigate human error The results of the review are provided directly to the IDS leadership and generally include a list of recommendations in the areas of improved legal agreements, leveraging privacy enhancing technologies to mitigate or reduce errors, procedural/policy gap identification and/or architectural recommendations to improve the security.

Exploring methods for generating synthetic data in Scotland to improve access to public sector data for research

Objective and ApproachWe are working to improve the economic, social and environmental wellbeing in Scotland by enabling access to and linkage of public-sector data for research in the public good. Our objective was to examine how synthetic data can support this aim by allowing researchers to test approaches and write code while awaiting the necessary permissions to use the real data. We have explored generating synthetic data from metadata and from actual data. ResultsThrough engagement, researchers expressed interest in low fidelity data, which minimises privacy and information governance risks. We have created low fidelity data from information provided by data controllers in publicly-available data dictionaries. We have also created more high-fidelity data based on actual data, using the synthpop tool. We are carrying out further work to engage with data controllers about the level of fidelity that is acceptable to them, as well as the information that needs to be included in the data dictionary to meet this in an automated way, and how we will share datasets with researchers. A key part of these considerations is the ability to link synthetic datasets and the impact that has on the fidelity of the data. Conclusions and ImplicationsWe will share our approaches, the limitations we have found in using metadata to create synthetic data, difficulties encountered in using the synthpop tool, engagement we have had from controllers and researchers, and the proposed way forward for synthetic data in Scotland.

Sensitive Data Flagging within Data Quality reports: R and Regex Integration for Effective Text Flagging in Large Datasets

ObjectiveWe’re introducing a comprehensive approach to enhance sensitive data flagging within our Data Quality (DQ) tool. Within de-identified health records, sensitive data such as user ids, post codes, phone numbers, and addresses pose significant privacy risks if exposed in their raw form. To address this challenge, we propose using all necessary regex patterns and free text field checks for sensitive data flagging with the objective of efficient detection within large datasets and ultimately reporting such data. ApproachA regex map has been developed along free text field thresholds, distribution counts and percentage checks to allow a generic approach. This ensures that any new regexes can be added to the flagging process without significant code changes. Once identified, sensitive data instances, including fields that contain free text are flagged and displayed in an html report that is human readable by DQ reviewers. ResultsThrough a combination of SQL and R regex based text processing, our approach allows a seamless identification and flagging of sensitive data within datasets. Sensitive data for large tables of +25 million records with over 50 columns are getting flagged in less than 150 seconds (approx. 2 minutes). ConclusionOur development offers a practical solution for sensitive data flagging in complex datasets where team members can implement robust sensitive data flagging by adding more regex formulas in the codebase. ImplicationsDQ reviewers can inspect the sensitive flagged fields in a human readable way, providing an extra measure of confidence when determining the quality of de-identified health records.

Establishment of an Australian National Master Linkage Key (NMLK) between AIHW Health Linkage spine (HLS) and Jurisdictional Master Linkage Keys (MLK) - NSW example

ObjectivesThis paper describes a Population Health Research Network project establishing the NMLK as a national linkage system utilising an enduring linkage between national and jurisdictional linkage spines. It used keys from existing enduring linkages within each participating jurisdictional Data Linkage Unit (DLU). The process is demonstrated with reference to an enduring map between AIHW Health Linkage Spine (HLS) and Centre for Health Record Linkage (CHeReL) MLK. ApproachThe NMLK has jurisdictional sections. The NSW section was established in four steps. CHeReL provided data from its regularly updated MLK for the NMLK project authorised by existing project ethics approval. AIHW linked NSW MLK subset to HLS and the de-identified linkage map and report were sent to CHeReL for review. CHeReL reviewed the results, particularly the additional duplicates identified through the linkage process. The agreed links form the AIHW-NSW section of NMLK quarantined for a year. The annual update, scheduled for later in 2024, will focus only on linking the new and unlinked individuals in both datasets. ResultsThe NSW cohort was supplied with almost 50 million records, comprising over 14 million unique person ids. Of these, 12 million were matched with reasonable confidence to HLS, or 84.4% of the total (or around 97% if we exclude NSW records with missing personal information and around 140,000 individuals out of scope of HLS.) Conclusion Establishment of the NMLK enabled effective production of high-quality enduring linkage maps used across multiple cross-jurisdictional projects while minimising the privacy risks.

Data vulnerability and privacy risk among hotel guests who share personal data

This research explores the guests’ perception toward data vulnerability and its impact on the privacy risk, privacy concern, and the perceived benefits of sharing personal data with hotels, which is in response to rising concerns about the personal data collection for service facilitation in hotels. A proposed research framework that explains the antecedents and outcomes of hotel guests’ privacy risk and privacy concern that is related to personal data sharing was examined using the generalized structured component analysis with measurement errors incorporated (GSCAM), the necessary condition analysis (NCA), and a fuzzy-set qualitative comparative analysis (fsQCA). The results expand the antecedents → privacy concerns → outcomes (APCO) model by determining a set of salient variables, such as data access, data breach vulnerabilities, privacy risk, and concern in regards to determining hotel guests’ perceived benefits of sharing personal data. This study provides valuable insights into privacy dynamics, enabling hotel professionals to develop targeted marketing strategies prioritizing both data utilization and guest privacy. It also suggests enhancing data security measures to mitigate concerns and foster trust among guests.

Efficiency Boosts in Human Mobility Data Privacy Risk Assessment: Advancements within the PRUDEnce Framework

With the exponential growth of mobility data generated by IoT, social networks, and mobile devices, there is a pressing need to address privacy concerns. Our work proposes methods to reduce the computation of privacy risk evaluation on mobility datasets, focusing on reducing background knowledge configurations and matching functions, and enhancing code performance. Leveraging the unique characteristics of trajectory data, we aim to minimize the size of combination sets and directly evaluate risk for trajectories with distinct values. Additionally, we optimize efficiency by storing essential information in memory to eliminate unnecessary computations. These approaches offer a more efficient and effective means of identifying and addressing privacy risks associated with diverse mobility datasets.

Social media and privacy concerns: exploring university student’s privacy concerns in TikTok platform in Vietnam

PurposeThis paper aims to investigate the extent of personal privacy concerns expressed by university students in Vietnam while using TikTok, the influence of peer interactions and social norms on privacy attitudes and behaviors and the strategies used by university students in Vietnam to mitigate privacy risks on TikTok.Design/methodology/approachA qualitative approach using semi-structured interviews was used to gather data on the following: the degree to which Vietnamese university students express concerns about their personal privacy while using TikTok; how peer interactions and social norms influence privacy attitudes and behaviors; and the strategies these students use to mitigate privacy risks on the platform.FindingsFindings indicate that although university students exhibit a relatively low level of concern regarding their personal privacy on TikTok, peer influences and societal norms significantly shape their attitudes and behaviors on the platform. Moreover, students use several strategies to mitigate privacy risks, such as selective content sharing and refraining from engaging with unknown links.Practical implicationsThese insights provide valuable implications for the development of targeted interventions and educational initiatives aimed at fostering a more privacy-conscious TikTok user community among university students in Vietnam.Originality/valueThis research fills a critical gap in the existing literature by focusing on the influence of social norms and peer influences on privacy attitudes and behaviors on social media platforms. While prior studies have explored various factors impacting privacy concerns on social media, the role of social dynamics has been understudied. Moreover, the research specifically addresses the lack of investigation into privacy concerns on TikTok, a platform rapidly gaining popularity among younger demographics.

Exploring Acceptability, Barriers, and Facilitators for Digital Health in Dermatology: Qualitative Focus Groups With Dermatologists, Nurses, and Patients.

Although several digital health interventions (DHIs) have shown promise in the care of skin diseases their uptake in Germany has been limited. To fully understand the reasons for the low uptake, an in-depth analysis of patients' and health care providers' barriers and facilitators in dermatology is needed. The objective of this study was to explore and compare attitudes, acceptability, barriers, and facilitators of patients, dermatologists, and nurses toward DHIs in dermatology. We conducted 6 web-based focus groups each with patients (n=34), dermatologists (n=30), and nurses (n=30) using a semistructured interview guide with short descriptions of DHIs described in the literature. A content analysis was performed using deductive constructs, following the unified theory of acceptance and use of technology framework, and inductive categories. Patients identified many positive performance expectancies, such as reduced travel times and improvement in follow-up appointments. Dermatologists also stated positive effects (eg, promotion of standardized care), but also negative implications of health care digitalization (eg, increased workload). All stakeholders reported that a DHI should bring additional value to all stakeholders. A lack of digital competence among patients was identified as the major barrier to adoption by all 3 groups. Nurses and dermatologists want apps that are easy to use and easy to implement into their daily routines. Trust in selected institutions, colleagues, and physicians was identified as a facilitator. Patients reported their dependence on the dermatologists' acceptance. All groups expressed concerns about data privacy risks and dermatologists stated insecurities toward data privacy laws. To ensure successful digitalization in dermatology, apps should be user-friendly, adapted to users' skill levels, and beneficial for all stakeholders. The incorporation of dermatologists' perspectives is especially important as their acceptance may impact use among patients and nurses. DHIs should ensure and be transparent about data privacy. The found barriers and facilitators can be used for implementation strategies.

Evaluating the disclosure risk of anonymized documents via a machine learning-based re-identification attack

The availability of textual data depicting human-centered features and behaviors is crucial for many data mining and machine learning tasks. However, data containing personal information should be anonymized prior making them available for secondary use. A variety of text anonymization methods have been proposed in the last years, which are standardly evaluated by comparing their outputs with human-based anonymizations. The residual disclosure risk is estimated with the recall metric, which quantifies the proportion of manually annotated re-identifying terms successfully detected by the anonymization algorithm. Nevertheless, recall is not a risk metric, which leads to several drawbacks. First, it requires a unique ground truth, and this does not hold for text anonymization, where several masking choices could be equally valid to prevent re-identification. Second, it relies on human judgements, which are inherently subjective and prone to errors. Finally, the recall metric weights terms uniformly, thereby ignoring the fact that the influence on the disclosure risk of some missed terms may be much larger than of others. To overcome these drawbacks, in this paper we propose a novel method to evaluate the disclosure risk of anonymized texts by means of an automated re-identification attack. We formalize the attack as a multi-class classification task and leverage state-of-the-art neural language models to aggregate the data sources that attackers may use to build the classifier. We illustrate the effectiveness of our method by assessing the disclosure risk of several methods for text anonymization under different attack configurations. Empirical results show substantial privacy risks for most existing anonymization methods.

Incorporating Privacy by Design Principles in the Modification of AI Systems in Preventing Breaches across Multiple Environments, Including Public Cloud, Private Cloud, and On-prem

The rapid integration of artificial intelligence (AI) across various sectors has significantly amplified privacy concerns, particularly with the growing reliance on cloud environments. Existing methods often fall short of effectively preventing privacy breaches due to inadequate risk assessment and mitigation strategies. These limitations highlight the necessity for more robust solutions, indicating the importance of Privacy by Design (PbD) principles. This study addresses these gaps by proposing a comprehensive approach to incorporating PbD principles into AI systems to prevent breaches across public, private, and on-prem environments. The proposed work utilizes logistic regression analysis to identify significant predictors of privacy breaches, revealing that both the environment (B = -1.142, p < .001) and severity of vulnerabilities (B = 0.932, p < .01) play crucial roles. Additionally, a strong positive correlation (r = 0.791) between breach detection rates and PbD effectiveness is observed, indicating the need for enhanced detection mechanisms. To support the empirical findings, this study also reviews existing case studies. It conducts a thematic analysis to provide a deeper understanding of the practical challenges and solutions associated with PbD implementation, particularly in healthcare and smart city applications. These analyses serve to supplement the empirical evidence and demonstrate the effectiveness of PbD over other existing methods. The study concludes that implementing PbD principles is critical for achieving robust privacy protection, and the study recommends prioritizing advanced breach detection mechanisms, comprehensive privacy impact assessments, continuous stakeholder engagement, and investment in privacy-enhancing technologies to address privacy risks effectively.

Reliable generation of privacy-preserving synthetic electronic health record time series via diffusion models.

Electronic health records (EHRs) are rich sources of patient-level data, offering valuable resources for medical data analysis. However, privacy concerns often restrict access to EHRs, hindering downstream analysis. Current EHR deidentification methods are flawed and can lead to potential privacy leakage. Additionally, existing publicly available EHR databases are limited, preventing the advancement of medical research using EHR. This study aims to overcome these challenges by generating realistic and privacy-preserving synthetic EHRs time series efficiently. We introduce a new method for generating diverse and realistic synthetic EHR time series data using denoizing diffusion probabilistic models. We conducted experiments on 6 databases: Medical Information Mart for Intensive Care III and IV, the eICU Collaborative Research Database (eICU), and non-EHR datasets on Stocks and Energy. We compared our proposed method with 8 existing methods. Our results demonstrate that our approach significantly outperforms all existing methods in terms of data fidelity while requiring less training effort. Additionally, data generated by our method yield a lower discriminative accuracy compared to other baseline methods, indicating the proposed method can generate data with less privacy risk. The proposed model utilizes a mixed diffusion process to generate realistic synthetic EHR samples that protect patient privacy. This method could be useful in tackling data availability issues in the field of healthcare by reducing barrier to EHR access and supporting research in machine learning for health. The proposed diffusion model-based method can reliably and efficiently generate synthetic EHR time series, which facilitates the downstream medical data analysis. Our numerical results show the superiority of the proposed method over all other existing methods.

Chinese Consumers' Attitudes toward and Intentions to Continue Using Skill-Sharing Service Platforms.

The sharing economy has rapidly transformed traditional consumption patterns worldwide. The emergence of skill-sharing services-which allow individuals to share their skills, abilities, and time through online platforms-has recently garnered attention. In China, the demand for skill-sharing services continues to grow, as these services effectively meet consumer needs. To understand this growing demand, this study aims to explore users' attitudes and intentions toward the use of skill-sharing service platforms in the Chinese market. A survey was conducted that incorporated 500 Chinese users who had used skill-sharing service platforms over the previous three months. A total of 409 datasets were analyzed, using structural equation modeling and multiple group analysis, in AMOS 24.0. The results showed that performance expectancy, effort expectancy, social influence, facilitating conditions, and self-efficacy positively influenced users' attitudes toward skill-sharing services, while privacy, functionality, and safety risks negatively affected these attitudes. Users' attitudes toward skill-sharing services significantly enhanced their intentions to continue using them, with the level of trust playing a crucial moderating role between attitude and the intention to continue using these services. These findings provide a significant theoretical and practical foundation for the further development of skill-sharing service platforms, the optimization of marketing strategies, and future research.

Privacy‐preserving collaboration in blockchain‐enabled IoT: The synergy of modified homomorphic encryption and federated learning

SummaryThe proliferation of network devices capable of gathering, transmitting, and receiving data over the Internet has spurred the widespread adoption of Internet of Things (IoT) devices, particularly in resource‐oriented applications. Integrating blockchain, IoT, homomorphic encryption, and federated learning requires a balance between computational requirements and real‐time performance. Secure key management is crucial to maintain data privacy and integrity. Compliance with privacy regulations requires careful implementation of privacy‐preserving mechanisms in blockchain‐enabled IoT environments, which can be subjected to various attacks. Addressing these challenges requires interdisciplinary expertise, research, and innovation to develop more efficient and effective privacy‐preserving techniques tailored to the unique characteristics of such environments. This research introduces the Modified Homomorphic Encryption Federated‐based Adaptive Hybrid Dandelion Search (MHEF‐AHDS) algorithm as an effective framework to enhance security in blockchain‐enabled IoT systems. The amalgamation of Modified Homomorphic Encryption (MHE) and Federated Learning (FL) constitutes a potent alliance that addresses privacy concerns within collaborative and decentralized machine learning environments. This facilitates secure and adaptable data collaboration, effectively mitigating privacy risks associated with sensitive information. The integration of quantum machine learning into security applications presents an exciting opportunity for distinctive progress and innovation. Within this work, the Adaptive Hybrid Dandelion optimization algorithm, featuring an Initial search strategy, is employed for hyperparameter optimization thereby elevating the performances of the proposed MHEF‐AHDS method. Furthermore, the integration of smart contracts and Blockchain‐based IoT enhances the overall security of the proposed method. MHEF‐AHDS comprehensively tackles privacy, security, and scalability challenges through robust security measures and privacy enhancements. The performance evaluation of the MHEF‐AHDS method encompasses a thorough analysis based on key metrics such as throughput, latency, scalability, energy consumption, accuracy, precision, recall, and f1‐score. Comparative assessments against existing methods are conducted to gauge the effectiveness of the proposed method in addressing security, privacy, and scalability concerns.

Genomic Data Fusion using Paillier Cryptosystem

The proposed work performs secure data fusion using homomorphic encryption, specifically the Paillier cryptosystem. The Paillier cryptosystem allows computation to be performed on encrypted data without decrypting it first, thus ensuring the privacy and security of the computation. The experiment measures the algorithm's performance based on execution time, memory usage, security, accuracy, and scalability. The data-level Paillier cryptosystem approach is generally slower than the feature-level fusion method due to its more complex operations and computations. Scalability is limited by the time required for encryption, homomorphic addition, and decryption. Improving scalability can be achieved by parallelizing the encryption and decryption steps, optimizing the homomorphic addition algorithm, or using more efficient cryptographic primitives. The article compares the performance of the Paillier cryptosystem with differential privacy in terms of their advantages and disadvantages. By adopting a preemptive approach to data fusion security, healthcare organizations can minimize the risk of data breaches and protect patient privacy. Data fusion security is an important factor when dealing with medical records. In the field of medical records, data fusion refers to the method of combining multiple sources of data into a distinct record. This can include data from electronic health records (EHRs), medical imaging devices, wearable devices, and other sources. There are several security considerations that must be addressed when fusing data from multiple sources.