Privacy Policy Language Research Articles

Evaluating Quantized Llama 2 Models for IoT Privacy Policy Language Generation

Quantized large language models are large language models (LLMs) optimized for model size while preserving their efficacy. They can be executed on consumer-grade computers without the powerful features of dedicated servers needed to execute regular (non-quantized) LLMs. Because of their ability to summarize, answer questions, and provide insights, LLMs are being used to analyze large texts/documents. One of these types of large texts/documents are Internet of Things (IoT) privacy policies, which are documents specifying how smart home gadgets, health-monitoring wearables, and personal voice assistants (among others) collect and manage consumer/user data on behalf of Internet companies providing services. Even though privacy policies are important, they are difficult to comprehend due to their length and how they are written, which makes them attractive for analysis using LLMs. This study evaluates how quantized LLMs are modeling the language of privacy policies to be potentially used to transform IoT privacy policies into simpler, more usable formats, thus aiding comprehension. While the long-term goal is to achieve this usable transformation, our work focuses on evaluating quantized LLM models used for IoT privacy policy language. Particularly, we study 4-bit, 5-bit, and 8-bit quantized versions of the large language model Meta AI version 2 (Llama 2) and the base Llama 2 model (zero-shot, without fine-tuning) under different metrics and prompts to determine how well these quantized versions model the language of IoT privacy policy documents by completing and generating privacy policy text.

PrivData Network: A Privacy-Preserving On-Chain Data Factory and Trading Market

Privacy concerns often raise when sensitive data are collected, traded, and processed. The data owner typically loses her ultimate control of the data after data-outsourcing. In this work, we present the PrivData Network – a community-controlled privacy-preserving data factory and trading market. It can be viewed as a standalone data ecosystem that enables privacy-preserving data-driven workflows in a controlled environment for orchestrating and automating data movement and data transformation. In particular, we design a data encapsulation mechanism with privacy assurance, which can guarantee data privacy, usage policy compliance and metadata validity. We also design a privacy policy language and utilize a static analysis library that transfers the program to the defined policy language. To ensure the correctness of data processing, we propose a publicly verifiable secure multiparty computation protocol for mixed circuits, which guarantees the output correctness even if all parties are corrupted. Its online efficiency is comparable to conventional semi-honest secret-sharing-based MPC schemes. Finally, we implemented a prototype of our system in C++ and benchmark it on various tasks, such as biometric matching, logistic regression, and decision trees, etc.

DataProVe: Fully Automated Conformance Verification Between Data Protection Policies and System Architectures

Abstract Privacy and data protection by design are relevant parts of the General Data Protection Regulation (GDPR), in which businesses and organisations are encouraged to implement measures at an early stage of the system design phase to fulfil data protection requirements. This paper addresses the policy and system architecture design and propose two variants of privacy policy language and architecture description language, respectively, for specifying and verifying data protection and privacy requirements. In addition, we develop a fully automated algorithm based on logic, for verifying three types of conformance relations (privacy, data protection, and functional conformance) between a policy and an architecture specified in our languages’ variants. Compared to related works, this approach supports a more systematic and fine-grained analysis of the privacy, data protection, and functional properties of a system. Our theoretical methods are then implemented as a software tool called DataProVe and its feasibility is demonstrated based on the centralised and decentralised approaches of COVID-19 contact tracing applications.

Privacy Policy Analysis of Banks and Mobile Money Services in the Middle East

Privacy compliance of the Middle East’s financial sector has been relatively unexplored. This paper evaluates the privacy compliance and readability of privacy statements for top banks and mobile money services in the Middle East. Our analysis shows that, overall, Middle Eastern banks have better privacy policy availability and language distribution, and are more privacy compliant compared to mobile money services. However, both the banks and mobile money services need to improve (1) compliance with the principles of children/adolescent’s data protection, accountability and enforcement, and data minimization/retention, and (2) privacy statement texts to be comprehensible for a reader with ~8 years of education or less.

Privacy in Gaming

Video game platforms and business models are increasingly built on collection, use, and sharing of personal information for purposes of both functionality and revenue. This paper examines privacy issues and explores data practices, technical specifications, and policy statements of the most popular games and gaming platforms to provide an overview of the current privacy legal landscape for mobile gaming, console gaming, and virtual reality devices. The research observes how modern gaming aligns with information privacy notions and norms and how data practices and technologies specific to gaming may affect users and, in particular, child gamers. After objectively selecting and analyzing major players in gaming, the research notes the many different ways that game companies collect data from users, including through cameras, sensors, microphones, and other hardware, through platform features for social interaction and user-generated content, and by means of tracking technologies like cookies and beacons. The paper also notes how location and biometric data are collected routinely through game platforms and explores issues specific to mobile gaming and pairing with smartphones and other external hardware devices. The paper concludes that transparency as to gaming companies’ data practices could be much improved, especially regarding sharing with third party affiliates. In addition, the research considers how children’s privacy may be particularly affected while gaming, determining that special attention should be paid to user control mechanisms and privacy settings within games and platforms, that social media and other interactive features create unique privacy and safety concerns for children which require gamer and parent education, and that privacy policy language is often incongruent with age ratings advertised to children and parents. To contribute additional research value and resources, the paper attaches a comprehensive set of appendices, on which the research conclusions are in part based, detailing the technical specifications and privacy policy statements of popular games and gaming platforms for mobile gaming, console gaming, and virtual reality devices.

Privacy policies verification in composite services using OWL

Privacy has been an important issue for online services collecting customer data. P3P is a privacy policy language with a fixed vocabulary to express privacy practices of online services. The matching between the privacy practices (P3P policies) and users' privacy preferences facilitates the users to be aware of services' usage of their data. However, the change from single to composite online services raises more privacy concern due to the increasing amount of user data being collected, stored and shared. This change impacts on P3P since it was designed from a single service perspective. In addition, P3P allows the specification of policies containing semantic inconsistencies. In this paper, we extend P3P to be suitable for composite services and propose a formal semantics for P3P using OWL to facilitate reasoning about semantic ambiguities in P3P policies. The constraints defined in our ontology are used to verify potential semantic inconsistencies and to check for conflicts occurring from P3P policies of service members. We have implemented a P3P verification tool and verified five hundred P3P policies collected from actual websites. The verification result shows that more than half of these P3P policies contain conflicts.

Resolving Conflicting Privacy Policies in M-health based on Prioritization

Mobile health has recently gained a lot of attention. Biological, environmental and behavioral data collected from mobile devices can be analyzed and transmitted directly to the person, family or health professionals for immediate and individualized care. However, due to multiplicity of mobile applications and the heterogeneity of actors involved in patient care, conflicts among the privacy policies defined by the different actors can take place. Thus, we present in this paper an approach to resolve the problem of conflicting privacy policies in e-health/m-health environments using AHP (Analytic Hierarchy Process) prioritization technique. Conflicts detection and resolution are facilitated by the adoption of S4P formal privacy policy language used as a standardized language. Finally, a case study is suggested to illustrate how our solution can be applied to resolve such conflicts.

Is Privacy Policy Language Irrelevant to Consumers?

AbstractThis article reports the results of two experiments in which large, census-weighted samples of Americans read short excerpts from Facebook’s, Yahoo’s, and Google’s privacy policies, which are at issue in high-stakes privacy class-action lawsuits. Subjects were randomly assigned to read language from either vague policies, some of which had been adjudicated insufficient to notify consumers about the companies’ practices, or explicit policies. Though many experimental subjects read these privacy policy excerpts closely, subjects who saw the explicit policies did not differ from those who saw vague policies in their assessment of whether their assent to the policies would permit the corporate practices at issue. Subjects generally stated that agreement to either vague or explicit language authorized companies to collect or use their personal information, even though consumers regarded these corporate practices as intrusive. These experiments show that courts and laypeople can understand the same priv...

Disagreeable Privacy Policies: Mismatches between Meaning and Users’ Understanding

Privacy policies are verbose, difficult to understand, take too long to read, and may be the least-read items on most websites even as users express growing concerns about information collection practices. For all their faults, though, privacy policies remain the single most important source of information for users to attempt to learn how companies collect, use, and share data. Likewise, these policies form the basis for the selfregulatory notice and choice framework that is designed and promoted as a replacement for regulation. The underlying value and legitimacy of notice and choice depends, however, on the ability of users to understand privacy policies. This paper investigates the differences in interpretation among expert, knowledgeable, and typical users and explores whether these groups can understand the practices described in privacy policies at a level sufficient to support rational decision-making. This paper seeks © 2015 Joel R. Reidenberg, Travis Breaux, Lorrie Faith Cranor, Brian French, Amanda Grannis, James T. Graves, Fei Liu, Aleecia McDonald, Thomas B. Norton, Rohan Ramanath, N. Cameron Russell, Norman Sadeh and Florian Schaub. † For their comments on this study, the authors would like to acknowledge and thank Alessandro Acquisti, Noah A. Smith, and Shomir Wilson, and the participants at the 2014 TPRC 42nd Research Conference on Communication, Information and Internet Policy. Funding for this project was provided, in part, by the National Science Foundation under its Secure and Trustworthy Computing (SaTC) initiative grants 1330596, 1330214, and 1330141 for “TWC SBE: Option: Frontier: Collaborative: Towards Effective Web Privacy Notice and Choice: A Multi-Disciplinary Prospective” and by a Fordham Law School Faculty Research Grant. †† Respectively, Stanley D. and Nikki Waxberg Chair and Professor of Law, Fordham University; Assistant Professor of Computer Science, Carnegie Mellon University; Professor of Computer Science and Engineering & Public Policy, Carnegie Mellon University: Senior Research Programmer, Carnegie Mellon University; Research Fellow, Fordham Center on Law and Information Policy; Ph.D Candidate (Engineering and Public Policy) Carnegie Mellon University; Ph.D Candidate (Computer Science), Carnegie Mellon University; Director of Privacy, Stanford Center for Internet & Society; Privacy Fellow, Fordham Center on Law and Information Policy; Masters Candidate (Computer Science), Carnegie Mellon University; Executive Director, Fordham Center on Law and Information Policy; Professor of Computer Science, Carnegie Mellon University; Postdoctoral Fellow (Computer Science), Carnegie Mellon University. 40 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 30:1 to fill an important gap in the understanding of privacy policies through primary research on user interpretation and to inform the development of technologies combining natural language processing, machine learning, and crowdsourcing for policy interpretation and summarization. For this research, we recruited a group of law and public policy graduate students at Fordham University, Carnegie Mellon University, and the University of Pittsburgh (“knowledgeable users”) and presented these law and policy researchers with a set of privacy policies from companies in the e-commerce and news and entertainment industries. We asked them nine basic questions about the policies’ statements regarding data collection, data use, and retention. We then presented the same set of policies to a group of privacy experts and to a group of crowd workers representing typical Internet users. The findings show areas of common understanding across all groups for certain data collection and deletion practices, but also demonstrate very important discrepancies in the interpretation of privacy policy language, particularly with respect to data sharing. The discordant interpretations arose both within groups and between the experts and the two other groups. The presence of these significant discrepancies has critical implications. First, the common understandings of some attributes of described data practices mean that semiautomated extraction of meaning from website privacy policies may be able to assist typical users and improve the effectiveness of notice by conveying the true meaning of these policies. However, the disagreements among experts and disagreement between experts and the other groups reflect that ambiguous wording in typical privacy policies undermines the ability of privacy policies to effectively convey notice of data practices to the general public. The results of this research will, consequently, have significant policy implications for the construction of the notice and choice framework and for the U.S. reliance on this approach. The gap in interpretation indicates that privacy policies may be misleading the general public and that those policies could be considered legally unfair and deceptive. And, where websites are not effectively conveying privacy policies to consumers in a way that a “reasonable person” could, in fact, understand the policies, “notice and choice” fails as a framework. Such a failure has broad international implications since websites extend their reach beyond the United States.

HIPAA-Compliant Privacy Policy Language for e-Health Applications

Many e-health applications collect patient's health data and track how they are used by patients to enable and validate their effectiveness. Although e-health applications allow people to access healthcare services in easy and convenient way at the reduced cost, the lack of reliable and effective methods of privacy protection makes people hesitate to use e-health applications, and in turn, it becomes the biggest obstacle to the growth of e-Health applications. To overcome the drawback, in this paper, we first address the lack of consideration of health-related data on existing privacy policy languages and propose the HIPAA profile for existing languages, which contains the Health data schema and extensions to HIPAA-friendly policy languages. By using the HIPAA profile, e-health providers are able to specify HIPAA-compliant privacy policies and patients can express their privacy preferences on not only general usage and user data but also health-related data in detail. For better understanding, we present example policies for e-health applications and patients using the proposed profile.

Disagreeable Privacy Policies: Mismatches between Meaning and Userss Understanding

Privacy policies are verbose, difficult to understand, take too long to read, and may be the least-read items on most websites even as users express growing concerns about information collection practices. For all their faults, though, privacy policies remain the single most important source of information for users to attempt to learn how companies collect, use, and share data. Likewise, these policies form the basis for the self-regulatory notice and choice framework that is designed and promoted as a replacement for regulation. The underlying value and legitimacy of notice and choice depends, however, on the ability of users to understand privacy policies.This paper investigates the differences in interpretation among expert, knowledgeable, and typical users and explores whether those groups can understand the practices described in privacy policies at a level sufficient to support rational decision-making. The paper seeks to fill an important gap in the understanding of privacy policies through primary research on user interpretation and to inform the development of technologies combining natural language processing, machine learning and crowdsourcing for policy interpretation and summarization. For this research, we recruited a group of law and public policy graduate students at Fordham University, Carnegie Mellon University, and the University of Pittsburgh (“knowledgeable users”) and presented these law and policy researchers with a set of privacy policies from companies in the e-commerce and news & entertainment industries. We asked them nine basic questions about the policies’ statements regarding data collection, data use, and retention. We then presented the same set of policies to a group of privacy experts and to a group of non-expert users.The findings show areas of common understanding across all groups for certain data collection and deletion practices, but also demonstrate very important discrepancies in the interpretation of privacy policy language, particularly with respect to data sharing. The discordant interpretations arose both within groups and between the experts and the two other groups. The presence of these significant discrepancies has critical implications. First, the common understandings of some attributes of described data practices mean that semi-automated extraction of meaning from website privacy policies may be able to assist typical users and improve the effectiveness of notice by conveying the true meaning to users. However, the disagreements among experts and disagreement between experts and the other groups reflect that ambiguous wording in typical privacy policies undermines the ability of privacy policies to effectively convey notice of data practices to the general public. The results of this research will, consequently, have significant policy implications for the construction of the notice and choice framework and for the US reliance on this approach. The gap in interpretation indicates that privacy policies may be misleading the general public and that those policies could be considered legally unfair and deceptive. And, where websites are not effectively conveying privacy policies to consumers in a way that a “reasonable person” could, in fact, understand the policies, “notice and choice” fails as a framework. Such a failure has broad international implications since websites extend their reach beyond the United States.

Obtaining P3P privacy policies for composite services.

With the development of web services technology, web services have changed from single to composite services. Privacy protection in composite services is becoming an important issue. P3P (platform for privacy preferences) is a privacy policy language which was designed for single web services. It enables service providers to express how they will deal with the privacy information of service consumers. In order to solve the problem that P3P cannot be applied to composite services directly, we propose a method to obtain P3P privacy policies for composite services. In this method, we present the definitions of Purpose, Recipient, and Retention elements as well as Optional and Required attributes for P3P policies of composite services. We also provide an instantiation to illustrate the feasibility of the method.

Understanding privacy policies

Their is growing recognition that users of web-based systems want to understand, if not control, what customer's data is stored by whom, for what purpose, for what duration, and with whom it is shared. We inform current language-based privacy efforts with an empirical study of P3P--the W3C domain-specific language for privacy policies. We use methods of software language engineering to study usage profiles, correctness of policies, metrics, cloning, and language extensions. The study supports the conclusion that P3P's approach to policy validation is too weak to ensure correct use of the language. The study also discovers common, dominating policies, which may suggest a simpler approach to web privacy. Further, the study investigates a range of metrics for policies in an attempt to discover particularly interesting or complex policies. Finally, the study also attempts to discover symptoms of the need for extending the P3P language, but the found results are not conclusive here.