8-533 / 8-733 / 19-608 / 95-818: Privacy Policy, Law, and Technology

Semester Project

All students in this course will be required to complete a project that they work on throughout the semester. Students will work in small groups of three to six students. Expectations about the size of the project will scale with the size of the group.

Schedule

September 15 - Project assignment discussed in class
September 22 - Project selection survey due before class
October 6 - One-paragraph project description due (5 points)
October 27 - Project proposal due (10 points)
December 1 - Draft paper due (10 points)
December 8 - Poster fair (15 points)
December 16, noon - Final paper due (60 points)

The various project assignments due before the final paper are designed to make sure you are making progress on your project throughout the semester and to give you opportunities to get feedback on your work along the way. Your draft paper will be graded for completeness, not content. For example, you will receive full credit for your draft paper if it has all the expected components and it appears that you put some effort into your draft, even if the content is unpolished. However, if your draft is missing an essential component (for example, a bibliography), you will not receive full credit. You will also lose points for submitting project components late. All project-related assignments will be graded within one week if they are submitted on time. You may also submit these assignments early for early feedback.

Deliverables

One-paragraph Project Description

Turn in a one-paragraph description of the project you intend to complete. Please submit your one-paragraph description on Blackboard and put "project description NAME1 NAME2" in the header or first line of the document (where NAME1 and NAME2 etc. are the names of the people proposing this project). If you want early feedback or help deciding between a few project ideas, please discuss with the professor or TA. Only one submission per group is required.

Project Proposal

The project proposal should include:

You might think of the project proposal as being similar to a grant proposal (without the need to fill out government forms or prepare a budget request). In the process of preparing this proposal you should conduct a literature review so that you can cite the relevant related work in your proposal. Besides being a graded assignment, the project proposal serves as a way for you to organize your thoughts about how to proceed with your semester project and to communicate them to your instructor. You will receive feedback on your proposal that may result in some changes to your project plans.

Writing quality (grammar, spelling, clarity, etc.) will be taken into account in your grade.

Please submit your project proposal in PDF format on Blackboard. Only one submission per group is required.

Draft Paper

Your draft paper should be a nearly complete version of your final project report. It should include an updated version of the literature review, and background and motivation from your project proposal. Please leave place holders for anything that is still incomplete (such as results from your data) and explain briefly what you expect to add in the final paper. If you developed software or created something as part of the project, please provide screen shots, a link to a demo, or other information so that the instructor can give you feedback on that part as well. Your electronic submission should be a PDF file submitted on Blackboard. Only one submission per group is required.

Final Paper

Your project report should document the work you have done on your project. It should include an updated version of the literature review, and background and motivation from your project proposal. If your project primarily involved writing a paper, then your project report may be the only artifact you submit. On the other hand, if you developed software or created something as part of this project, you should submit whatever you created in addition to the report. In the latter case, the report should document what you did and may include information about obstacles you encountered, testing and evaluation, design rationale, etc., as appropriate. Please consult with the instructor about what should be included in your report if you have any doubts.

Students enrolled in the 12-unit versions of this course are expected to write up their report in a format suitable as a conference paper submission.

Because of all the opportunities you have to get feedback on your project during the semester, the final paper and poster presentation will be graded with fairly high standards. What we will be looking for depends a lot on the particular project you choose. Here are some things we will be looking for in most papers.

Please submit your final paper in PDF format on Blackboard. Only one submission per group is required.

Poster

A poster session (open to the public) will be scheduled during the last week of classes. You should prepare a poster that provides an overview of your project. A 32x40 inch foam core board and easel will be provided to each student. I will also provide thumb tacks, construction paper, glue sticks and other supplies. You may prepare your poster as a set of up to 9 8.5x11 sheets of paper or print it as a single sheet. SCS provides a large format poster printer by the SCS computing facilities help desk. More details about the poster session will be provided in class.

Your poster grade will be based on the content of your poster, the visual presentation, your oral presentation, and your ability to answer questions. Be prepared to give a three-minute presentation to your instructor or other poster evaluator and answer their questions.

Project ideas

The following are a list of suggested projects. Students may select one of these projects or develop their own project idea in consultation with the instructor.

Privacy implications of social bots

Bots on the Internet can range from spambots to denial-of-service bots and often have some sort of financial or malicious motive. One of the less understood types of bots are the seemingly benign bots on social media sites (e.g., Twitter, Instagram) that automatically follow users and like or comment on their pictures in order to elicit the same in return to grow their network and social impact [1]. Often these bots are real users making use of an automatic bot service; other times the accounts are inhabited only by bots. When users allow themselves to be followed by "users" like these or follow such "users" in return, they open themselves up to privacy violations, harm, or data theft. This project will study: 1) to what extent users interact with unknown users on social media, specifically, potentially fake or harmful user accounts or bots; and 2) what the potential goals of these kinds of bots are and whether they achieve them.

[1] https://arxiv.org/pdf/1407.5225.pdf

Explore the accuracy of online profiling

While it is common knowledge that users are tracked across websites to create profiles for behavioral advertising, little is known about the profiles themselves and their accuracy. Students working on this project will study profile transparency pages like Google's Ad Settings page or the bluekai Registry. They will develop an online survey to ask people whether the information presented on these pages is correct and collect their impressions (e.g., whether they are surprised, shocked). You will need IRB approval for this project--plan to get it early.

Publicly shared photos and privacy implications

Publicly shared photos on websites such as pinterest, flickr, facebook, and twitter can reveal large amounts of information about the person who uploaded the image and other people in the photo. For example, EXIF data or location-inferring machine learning algorithms can leak information about participants' location [1, 2]. Such unintended leaks can have many consequences for users' privacy. For example, over time, this information may reveal behavioral patterns that can be used to predict future location. If Jordan takes a photo in the same restaurant every Friday, an observer could infer their future location. At the same time, this publicly available information could reveal undesirable things about past behavior. For example, Jordan may have told Hayden they would be out of town while publicly posted photos suggest otherwise. Other than location data, the identity of people in the photos may suggest privacy sensitive information, like the strength of social ties.

Potential projects in the area include: (1) Scraping publicly available pictures on the Internet to analyze EXIF or other location-inferable information and measuring the incidence of possible privacy violations. (2) Experimenting with machine learning algorithms for location inference and training on publicly available information; measuring the extent to which location can be inferred both in general and specifically when this is inconsistent with what people chose to reveal. (3) Analyzing the occurrence of transitive information disclosure of photos. For example, attempting to infer the strength of social ties of users. (4) Empirically identifying privacy leaks via photo sharing. (5) Predictive location modeling based on information leaks from public photos. We also encourage you to explore your own direction in this area.

[1] https://en.wikipedia.org/wiki/Exif#Privacy_and_security

[2] https://www.technologyreview.com/s/600889/google-unveils-neural-network-with-superhuman-ability-to-determine-the-location-of-almost/

Integrating privacy information into online forms

Develop a browser extension that detects when websites present a form to users and identifies which data types are asked from the user. For each identified field the browser extension should determine whether the website's privacy policy specifies how this data will be used, e.g., for what purposes and whether it is shared with third parties (if yes, for what reasons). For the project it is sufficient to show information for a fixed set of websites of which privacy policies have been analyzed beforehand. The focus is on a proof-of-concept integration of privacy information into the website's form in a usable and meaningful manner. The project might include a small user study to evaluate the effectiveness of the developed approach.

Developing privacy personas

Design and conduct a survey or interview study to gain insights on motivations of individuals seeking to protect their privacy. What do they want to protect from whom and why? From the study results derive a set of "privacy personas," i.e., prototypical characterizations of users with specific privacy motivations and associated needs for privacy-enhancing technologies. This project will require IRB approval, so get that early!

Standardized privacy notice for wearable and internet of things (IoT) devices

Standardized notice formats have been proposed and are in use in different domains (e.g., privacy nutrition labels for websites, Gramm-Leach-Bliley Act [GLBA] privacy notice requirements for financial institutions). Wearable devices, such as activity trackers, smart clothing or Google Glass, as well as IoT devices, such as smart thermostats or connected cameras, pose novel privacy challenges due to their continuous collection of data and physical proximity to the user's body. The goal of this project is to analyze requirements for a standardized notice format for such devices. What data practices should be emphasized? What user requirements and concerns need to be taken into account? The team should focus either on wearables or IoT devices. The project would include a review of the existing literature on privacy in this area (see A Design Space for Effective Privacy Notices to get started), interviews or surveys with wearable/IoT users and non-users, and a technical analysis of wearable/IoT data practices in order to derive design recommendations for a standardized short-form privacy notice for wearables or IoT devices. This project will require IRB approval, so get that early!

Improving data security breach notices

The state of California has a law that requires companies that experience a breach of their data security to provide a notice to the state and those affected that the breach has occurred. These breach notices are required to have a format that makes them easier to read. (The format is described in this amendment: here.) Notices can be formatted in two different ways. For this project, you will perform user testing, through a survey or interviews, to determine the effectiveness and usability of the two new formats as compared to what was previously used, examples of which can be found here. This project will require IRB approval, so get that early!

Temporary Messaging Apps and Privacy

Temporary messaging features, such as those offered by Snapchat and Instagram, are a new trend in social media. However, much of the privacy offered for these features rely on the messages &disappearing after a short time. For this project, conduct a study of users of temporary messaging services, such as Snapchat or Instagram stories, to determine their strategies for protecting their privacy. Do they self-censor, setup multiple accounts, block people, etc.? What strategies are most popular? What privacy threats do people believe they are protecting against? What privacy threats do they feel they have not adequately protected against? You might use interviews, focus groups, or surveys for this project. You will need IRB approval for this project -- plan to get it early.

International Privacy Survey

For decades, a small number of researchers have aimed to understand the international dimensions of privacy. For instance, in what ways do privacy norms compare and contrast across cultures? What types of behaviors are considered private in some countries, yet public in others? Are privacy concerns universal, or are some privacy concerns isolated to particular countries? Does the conception of privacy itself differ across countries? Many past studies of cultural differences in privacy have suffered from small or biased samples, surveys that were only available in English, or surveys that were restricted to only a small number of countries. Design and conduct an international privacy survey that investigates some of these questions using Amazon's Mechanical Turk or another popular crowdsourcing site. You will need IRB approval for this project--plan to get it early.

Privacy Notice -- In What Language?

The Internet has made the world smaller. Users from hundreds of different countries, speaking hundreds of different languages, access globally popular websites like Google, Facebook, Wikipedia, and Pinterest. However, members of Professor Cranor's CUPS lab have recently shown that even if a site is offered in a particular language, critical privacy information is not necessarily available at all in that language. The Dutch Data-Protection Authority took action against a company in part for failing to translate privacy information to Dutch for users in the Netherlands. In this project, you will systematically identify global websites or third-party advertising companies that regularly collect information from users around the world and quantify the extent to which they translate privacy-critical information into other languages.

Privacy software development

Design and implement a privacy-related software tool that offers functionality or features that are different from the other tools currently available. You might develop a stand-alone tool or develop a module for another piece of software, for example Mozilla. Depending on the scope of what you have in mind, it may not be feasible to implement your entire design during this semester, in which case you should implement one component of the design and document the rest of the design, perhaps also implementing a mocked up user interface. Your report should explain the rationale behind your design, the types of privacy protections this software offers, who would be interested in using it, and how it differs from other software currently available.

Web measurement study

Conduct a web measurement study that provides data on the prevalence of tracking, advertising, or privacy leakages. You might study whether Do Not Track is effective in preventing tracking or behavioral ads, or whether there is racial bias or price discrimination in e-commerce sites or ads. This could be a replication of previous work using more recent technologies, or focus on a new area.

Privacy and robots

What privacy issues are raised when robots operate in close proximity to people? How close is too close for drones to fly before they raise privacy concerns and heighten levels of fear and anxiety? Aerial robots are becoming increasingly capable of autonomously operating near people and animals. As the technology evolves (and safety guarantees increase), it is likely policies will allow systems to operate closer to people. As expected, one can empirically observe that these systems yield a heightened level of fear and anxiety when operating proximal to people. See here and here . And it turns out bears don't like drones either.. The evaluation of policies leading to FAA guidelines is based almost entirely on spatial safety. However, proximal safe operation will evolve. There may be some benefit of understanding the impact of the technology on the individual. This project will involve conducting background research on policy, precedents, related-lateral examples of psychological impact, and experimental evaluation (survey or experiment with people and drones) to establish correspondent empirical/primitive results suggesting potential guidelines and how these values change between individuals. The project could focus on self-driving cars in addition to or instead of drones. You will need IRB approval for this project--plan to get it early.

Ad and tracker blocking

The Future of Privacy Forum is interested in gaining a better understanding of why people use ad and tracker blocking tools and understanding the implications of those tools for websites. Are people mostly concerned about the ads or the tracking? How effective are the tools as detecting ads or tracking? What websites break when people use these tools? Students working on this project could conduct a survey of tool users and/or conduct a study in which they use adblock+, Ghostery, and other tools -- such as Safari Content Blocking or Opera automatic ad blocking -- and visit a large number of sites to determine how these tools impact the user experience at these sites. Note, that some sites have agreements with agreements so they won't have their ads blocked.