Compromised Privacy Research Articles

Honor the Contract? Effects of Algorithmic Recommendation System Features on Perceived Benefits, Privacy Risk, and Continuance Intention to Use TikTok

Algorithmic recommendations, characterized by features including accuracy, familiarity, novelty, and transparency, have been used in TikTok for a long time. However, the impact of these features on users’ privacy calculus and continuance usage intention remains unclear. Survey data from 625 Chinese users was analyzed using the least-squares partial structural equation modelling. Results showed that four features positively affect users’ perceived benefits; perceived familiarity positively affects privacy risk; perceived effectiveness of privacy policy moderates the relationship between perceived accuracy and privacy risk. Moreover, perceived benefits positively influence continuance intention to use, whereas privacy risk has a negative impact. This study shed light on why users are willing to compromise privacy for the benefits originating from features of algorithmic recommendations. Furthermore, this study provided valuable insights for platforms and practitioners concerning the design and development of effective algorithmic recommendations, as well as improving the effectiveness of privacy policy.

Enhancing Security and Privacy in Large Language Model-based Approaches: A Comprehensive Investigation

During the last five years, there have been significant developments in the field of Natural Language Processing including the deployment of advanced large language models such as ChatGPT, Bard and Llama. These large language models are helpful in generating text and designing content and they have several applications in various industries. However, they can memorize and reveal malicious content and personal information from their training dataset which also includes an enormous amount of data from the internet. As a result, it can lead to compromised privacy and security challenges for users who have their personal information available on the internet directly or through third parties. To address this issue, the proposed research work conducts a thorough investigation of these challenges and puts forward a prompt designing-based solution. In this method, we build a customized training dataset to fine-tune a pre-trained model (Llama-2) to produce a harmless response ‘I can’t provide you with this information’ to prompts seeking to extract personal information and malicious content from LLMs. Experimental results reveal that the proposed work achieves an accuracy of 63% with a precision score of 0.706 and a recall score of 0.571. The work ensures almost no leakage of private information and strengthens the LLM model against extraction attacks.

A Survey Study on IoT Application and its Attacks

The Internet of Things (IoT) has witnessed exponential growth in recent years, with billions of interconnected devices seamlessly communicating and exchanging data. While IoT offers unprecedented opportunities for innovation and convenience, it also introduces a plethora of security challenges. This abstract provides an overview of the key security concerns in IoT and explores emerging solutions to address them. Security in IoT is paramount due to the potential consequences of breaches. IoT devices are often resource-constrained and lack robust security features. Malicious actors can exploit this vulnerability to gain unauthorized access, compromise privacy, launch cyberattacks, and disrupt critical services. Therefore, securing IoT ecosystems is imperative for the continued growth and adoption of IoT technologies.

A dynamic anonymous authentication scheme with trusted fog computing in V2G networks

As an important component of smart grid, vehicle-to-grid (V2G) is helpful to improve grid stability, save on generation costs, and reduce energy consumption. However, the existing V2G authentication schemes focus on the trustworthiness of electric vehicles (EVs), instead of the trustworthiness of fog servers (FSs). For unsuspecting EVs, malicious FSs may steal their sensitive data and compromise privacy. We propose a dynamic anonymous authentication scheme with trusted fog computing in V2G networks. In our scheme, to ensure the trustworthiness of the FSs in V2G networks, a collaborative trust evaluation model among FSs is designed. The trust value of FSs can be calculated by comprehensively analyzing the timeliness, authenticity, and execution of FSs' behavior in the collaboration process. Meanwhile, trust evaluation is incorporated into the dynamic anonymous authentication scheme against malicious FSs and ensure a trusted environment of EVs in the authentication process. During the authentication process, to avoid the key escrow and computational overhead of FSs, the certificateless mechanism and aggregated signatures of non-linear pairs are used. We use a combination of long-term anonymity (LID) and temporary anonymity (TID) to ensure the anonymity and the signatures freshness of EVs. Finally, security proofs based on random oracle models and non-formal security analysis show that our scheme satisfies seven security requirements and resists five common attacks. The simulation results show that, in comparison with the five reference schemes, our scheme reduces average computational and communication overheads by 80.05% and 47.68%, respectively.

Compromise privacy in large-batch Federated Learning via model poisoning

Federated Learning (FL) is a distributed learning paradigm in which clients collaboratively train a global model with shared gradients while preserving the privacy of local data. Recent researches found that an adversary can reveal private local data with gradient inversion attacks. However, setting a large batchsize in local training can defense against these attacks effectively by confusing gradients computed on private data. Although some advanced methods have been proposed to improve the performance of gradient inversion attacks on the large-batch training, they are limited to specific model architectures (e.g., fully connected neural networks (FCNNs) with ReLU layers). To address these problems, we propose a novel gradient inversion attack to compromise privacy in large-batch FL via model poisoning. We poison clients' models with malicious parameters, which are constructed purposely to mitigate the confusion of aggregated gradients. For FCNNs, the private data can be perfectly recovered by analyzing gradients of the first fully connected (FC) layer. For convolutional neural networks (CNNs), we extend our proposed method as a hybrid approach, consisting of the analytic method and the optimization-based method. We first recover the feature maps after the convolutional layers and then reconstruct private data by minimizing the loss of data-wise feature map matching. We demonstrate the effectiveness of our proposed method on four datasets and show that it outperforms previous methods for large-batch FL (e.g., 64, 128, 256) and models with different activation layers (e.g., ReLU, Sigmoid and Tanh).

A Hybrid Architecture Based on Blockchain to Ensure Security, Privacy, and Trust in IoT

IoT is an enormous network of interconnected devices that gather data and share it over the internet. Today, IoT suffers from malevolent behaviours and actions that aim to damage and compromise privacy and security. Moreover, these issues and the new IoT requirements are not effectively covered by current architectures. Thus, this paper is an extended work of the authors' previous paper, which suggested a new architecture adhering to the IoT needs that have evolved. It suggests a hybrid architecture that integrates our previous one with a Blockchain-based architecture, to combine their significant features. The hybrid architecture fulfils the emerged requirements of IoT and exploits the Blockchain to increase trust, security and privacy. Further, the hybrid architecture delivers a novel structure to build the policy header, which is an access control list used to regulate the local network and was inherited from the Blockchain-based architecture. The Cooja simulator is used to compare the two structures, revealing that the new structure can use 87% less storage and 83% less computing.

"Who is driving you home today?" Escort policies as a barrier to abortion access

Post-sedation escort policies are not evidence-based but traditional consensus recommendations made by professional societies. As people travel further for abortion care, escort policies are increasingly difficult to navigate and force people to delay care, compromise privacy, or undergo procedures without sedation. At worst, clinics may turn away people who present without an escort. Recent research shows that patients can be discharged safely after sedation using rideshare or transport services without a known escort. Updating escort policies lowers barriers to abortion and preserves autonomy, comfort, and choice.

Customer Purchase Journey, Privacy Choices, and Advertising Strategies

Advertising to a consumer moves her further along the purchase journey, and tracking the consumer's online activities enables the advertiser to infer her purchase journey stage and target her accordingly. However, many consumers dislike being tracked, and repeated advertising leads to ad wearout. Under these considerations, we investigate the impact of regulations that endow a consumer with the choice to opt in to have her activity be tracked or opt out of tracking. A consumer's opt-in decision balances two aspects of privacy: its intrinsic value (protect privacy for its own sake) and its instrumental value (compromise privacy if doing so leads to utility-enhancing outcomes). Consumers opt in or out of tracking under different conditions. For instance, when ad effectiveness is intermediate, such that repeated ads would be useful for the advertiser, consumers opt in. This is because tracking reveals the consumer's purchase journey stage, which increases advertising efficiency and thus reduces ad repetition than if the consumer had opted out. Due to the opt-in/-out choice, and compared to the pre-regulation benchmark of consumer data being tracked, consumer surplus weakly increases and the advertiser's and the ad network's profits weakly decrease. Interestingly, the advertiser's and the ad network's profits may decrease in ad effectiveness.

A Survey on Mobile Security Issues

In today`s world of advanced science the world is changing, so is our technology. Smart- phones have become crucial part of day to day life. The increased use of smart-phones has given rise to ways we could never think like easy accessibility to net, work or personal information. But with it comes to the risk of security issues like compromised privacy or unauthorized accesses to mobile phones etc, to name some the unique structure makes easy target for malware developers and Hackers. Android for example has a very few restrictions so there is very less security system for end users. So the first target should be to eliminate these security breaches and give users the smooth experience without worrying about these problems. The paper addresses the problems faced with mobile security and how to overcome them. The paper also presents a brief description about what kind of operating system is much more at risk of these attacks and, which is more prone to security breach. It also consists of what fields still need to be researched on as to move forward in better direction. Furthermore, it explains the working of wireless systems and the attacks on it by hackers and how can that be prevented. Lastly it also illustrates about the attacks by hackers, types of attacks done on mobile phones and other related wireless systems and how to prevent them, and what users can do to prevent themselves by these attacks.

Childbirth experiences and their derived meaning: a qualitative study among postnatal mothers in Mbale regional referral hospital, Uganda

BackgroundEvidence shows that negative childbirth experiences may lead to undesirable effects including failure to breastfeed, reduced love for the baby, emotional upsets, post-traumatic disorders and depression among mothers. Understanding childbirth experiences and their meaning could be important in planning individualized care for mothers. The purpose of this study was to explore childbirth experiences and their meaning among postnatal mothers.MethodsA phenomenological qualitative study was conducted at Mbale Regional Referral Hospital among 25 postnatal mothers within two months after birth using semi-structured interviews and focus group discussions and data was thematically analyzed.ResultsThe severity, duration and patterns of labour pains were a major concern by almost all women. Women had divergent feelings of yes and no need of biomedical pain relief administration during childbirth. Mothers were socially orientated to regard labour pains as a normal phenomenon regardless of their nature. The health providers’ attitudes, care and support gave positive and negative birth experiences. The Physical and psychosocial support provided comfort, consolation and encouragement to the mothers while inappropriate care, poor communication and compromised privacy contributed to the mothers’ negative childbirth experiences. The type of birth affected the interpretations of the birth experiences. Women who gave birth vaginally, thought they were strong and brave, determined and self-confident; and were respected by members of their communities. On the contrary, the women who gave birth by operation were culturally considered bewitched, weak and failures.ConclusionChildbirth experiences were unique; elicited unique feelings, responses and challenges to individual mothers. The findings may be useful in designing interventions that focus on individualized care to meet individual needs and expectations of mothers during childbirth.

An IP Core Remote Anonymous Activation Protocol

The IP-based design of Systems-on-chip, both for FPGAs and ASICs, is enabling various new business models, where the essential aspect is the decoupling of system integration from the development of the individual cores. Such new scenarios also bring inherently new security challenges for the protection of the intellectual property. On the other hand, online tracking requires some exposure of the user's platform, which may result in compromised privacy. This work relies on a concept borrowed from the trusted computing area, Direct Anonymous Attestation, to enable remote IP licensing and activation mechanisms preserving the full anonymity of the end user, i.e., making it impossible for the Activation Server (AS) in charge of distributing the IP core license to track the user's behavior across the SoC lifecycle. The main findings of this work lie in the definition of an ad-hoc protocol, called Remote Anonymous Activation Protocol (RAAP), as well as a proof-of-concept implementation. The experimental results collected from the prototype confirm the computational feasibility of the proposed protocol on a representative example of a medium-end system-on-chip.

Sounding the Alarm on Digital Self-harm

Research Article| January 01 2018 Sounding the Alarm on Digital Self-harm AAP Grand Rounds (2018) 39 (1): 5. https://doi.org/10.1542/gr.39-1-5 Views Icon Views Article contents Figures & tables Video Audio Supplementary Data Peer Review Share Icon Share Facebook Twitter LinkedIn MailTo Tools Icon Tools Get Permissions Cite Icon Cite Search Site Citation Sounding the Alarm on Digital Self-harm. AAP Grand Rounds January 2018; 39 (1): 5. https://doi.org/10.1542/gr.39-1-5 Download citation file: Ris (Zotero) Reference Manager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search toolbar search search input Search input auto suggest filter your search All PublicationsAll JournalsAAP Grand RoundsPediatricsHospital PediatricsPediatrics In ReviewNeoReviewsAAP NewsAll AAP Sites Search Advanced Search Topics: self-injurious behavior, cyberbullying Source: Patchin JW, Hinduja S. Digital self-harm among adolescents. J Adolesc Health. 2017 Dec; 61(6): 761– 766; doi: https://doi.org/10.1016/j.jadohealth.2017.06.012Google Scholar Investigators from the University of Wisconsin-Eau Claire and Florida Atlantic University conducted a survey of US middle and high school students to quantify the proportion of students who engage in digital self-harm. Students were eligible if they were English-speaking and aged 12–17 years. Surveys were administered via email. Two survey questions were used to assess digital self-harm: (1) “in my lifetime, I have anonymously posted something online about myself that was mean,” and (2) “in my lifetime, I have anonymously cyberbullied myself online.” For those who answered affirmatively, the frequency of engaging in self-harm was assessed (once, a few times, many times) as were the motivations for doing so. Demographics and data on several covariates were also collected, such as sexual orientation, depressive symptoms, offline self-harm, drug use, and history of being a victim of school or cyberbullying. Investigators determined the number of respondents who engaged in digital self-harm and the frequency in which digital self-harm was used. Investigators used multivariable logistic regression to assess the independent association of digital self-harm with covariates after controlling for age, sex, and race. There were 5,593 students included in analysis, representing a 15% response rate. Most were white (66%), and half were female (49.9%). The proportion of respondents who “anonymously posted something online that was mean” about themselves or “anonymously cyberbullied” was 6.2% and 5.3%, respectively. Most respondents stated they engaged in this self-harm just once. Motivations for engaging in self-harm included self-hate, to be funny, looking for a reaction, and attention-seeking. Significantly more boys reported engaging in digital self-harm. In multivariable analyses, being non-heterosexual, a victim of school or cyberbullying, drug use, depressive symptoms, and engaging in offline self-harm were all significantly associated with engaging in digital self-harm. The investigators conclude that 1 in 20 middle and high school students engage in digital self-harm. Dr Wong has disclosed no financial relationship relevant to this commentary. This commentary does not contain a discussion of an unapproved/investigative use of a commercial product/device. When providing preventive care to adolescent patients, the potential harms of social media and screening for self-harm are increasingly prevalent and important issues to discuss. The potential harms of social media tend to be focused on risks posed by others online. Teenagers using social media, for example, are at risk of cyberbullying, which has been linked to multiple mental health issues, including attempted suicide (see AAP Grand Rounds, September 2012;28[3]).1 Other harms include compromised privacy and confidentiality, exposure to online predators, and misinformation. Like the rise in social media use and risks, the rate of self-harm among adolescents (eg, cutting) has also been increasing. Recent estimates suggest that >10% of adolescents report ever having self-harmed.2 The current investigators describe a more recent form of self-harm called digital self-harm. They define digital self-harm as “anonymous online posting, sending, or... You do not currently have access to this content.

Longitudinal Global Positioning System Travel Data and Breach of Privacy via Enhanced Spatial and Demographic Analysis

Longitudinal Global Positioning System (GPS) travel data provide a wealth of information related to travel behavior and on-road vehicle behavior that is very valuable to researchers. Sharing the data publicly allows researchers to explore the data and create new knowledge beyond the initial research objectives. However, if any data are to be used outside a secure server, the data must be processed in such a manner that ensures that the confidentiality of the data will not be breached. High-resolution GPS data (e.g., second-by-second speed and location information), when associated with the individual households or drivers, compromise privacy and have a significant potential to harm human subjects. This paper explores how data from the Commute Atlanta study in Georgia could be processed to make it useful to researchers while participants’ privacy is protected. The research developed and assessed methodologies designed to identify the individual participant's home location from processed data and then tested analytical data sets for breach of privacy. The research effort found that the home location could be identified to within reasonably small neighborhoods; when the household demographic information was included in the data sets (which was necessary for researchers), exact households could be identified. Although some new data-processing approaches might be used to eliminate privacy concerns, until such systems are developed and proved to be unbreachable through rigorous analysis, the Georgia Institute of Technology team has determined that researchers should access the high-resolution data in controlled secure labs and that the data sets should not be made public without additional efforts to ensure that home locations cannot be identified when external data sources are leveraged in the analyses.

Participatory Surveillance in the Intelligent Building

Introduction The rationale behind the so-called intelligent building1 is a vision of more flexible interaction with computers. It makes computers less visible, if not invisible, as they become part of the surroundings rather than being objects that need direct instructions through a formal interface. The intelligent building is conceived as a caring environment where computers adapt to human existence, rather than the other way around. However, this sympathetic idea raises privacy questions. For the intelligent building to be an adaptive, caring environment, it needs to generate and store quite a lot of information about the behaviors and lifestyles of the inhabitants. Basically, the perfect intelligent building needs to monitor everything that goes on within the walls of the house. The information generated, which when isolated may seem unimportant, flows together to form meaningful information images of the inhabitants. Furthermore, it seems likely and desirable that the intelligent building should be connected to the outside world via the Internet, in the same way that most digital devices are connected today. In this way, computation—which itself is almost invisible—has the potential to make home and work life visible to the world. As indicated, the discourse surrounding the intelligent building seems to have two contrasting visions that accompany this scenario: The dream of efficiency and care and an opposing nightmare vision of compromised privacy and control. These opposing visions are mirrored in surveillance studies, where care opposite control is recognized as the “Janus face” of surveillance2— although a more complex characterization can be deduced from the wide variety of domains studied, ranging from social control3 to empowerment and resistance,4 art and entertainment,5 and social practices.6 These opposing visions and discourses not only are relevant in relation to intelligent buildings, but are equally important when designing technology for various contexts and purposes where the same tensions surface (e.g., personalized search engines, GPS and camera surveillance in taxis, or adaptive learning environments) and thus have wider currency in design studies. In the following we argue that the two visions of the intelligent house share a focus on the built structure itself and, paradoxically, on the external stakeholders (e.g., caregivers, potential privacy invaders, etc.) rather than on the in habitants of the house.

Human Responsibility for Autonomous Agents

Automated or autonomous systems can sometimes fail harmlessly, but they can also destroy data, compromise privacy, and consume resources, such as bandwidth or server capacity. What's more troubling is that automated systems embedded in vital systems can cause financial losses, destruction of property, and loss of life. Controlling these dangers will increase trust while enabling broader use of these systems with higher degrees of safety. Obvious threats stem from design errors and software bugs, but we can't overlook mistaken assumptions by designers, unanticipated actions by humans, and interference from other computerized systems. This article is part of a special issue on Interacting with Autonomy.

Pediatric home healthcare: a paradox.

Although parents may welcome having their ill child cared for at home, they are not prepared to compromise privacy and family rituals, nor share control of their child. The purpose of this article is to provide a snapshot of problems that parents have encountered with pediatric home healthcare. Home care parents offer suggestions for other parents and home healthcare nurses and agencies, encouraging them to be proactive in preventing potential problems.