To resist government and corporate use of facial recognition to surveil users through their personal images, researchers have created privacy-enhancing image filters that use adversarial machine learning. These "subversive AI'' (SAI) image filters aim to defend users from facial recognition by distorting personal images in ways that are barely noticeable to humans but confusing to computer vision algorithms. SAI filters are limited, however, by the lack of rigorous user evaluation that assess their acceptability. We addressed this limitation by creating and validating a scale to measure user acceptance --- the SAIA-8. In a three-step process, we apply a mixed-methods approach that closely adhered to best practices for scale creation and validation in measurement theory. Initially, to understand the factors that influence user acceptance of SAI filter outputs, we interviewed 15 participants. Interviewees disliked extant SAI filter outputs because of a perceived lack of usefulness and conflicts with their desired self-presentation. Using insights and statements from the interviews, we generated 106 potential items for the scale. Employing an iterative refinement and validation process with 245 participants from Prolific, we arrived at the SAIA-8 scale: a set of eight items that capture user acceptability of privacy-enhancing perturbations to personal images, and that can aid in benchmarking and prioritizing user acceptability when developing and evaluating new SAI filters.

Many people are concerned about how their personal data is used for online behavioral advertising (OBA). Ad targeting explanations have been proposed as a way to reduce this concern by improving transparency. However, it is unclear when and why people might want ad targeting explanations. Without this insight, we run the risk of designing explanations that do not address real concerns. To bridge this gap, we conducted a four-week, mixed-methods field study with 60 participants to understand when and why people want targeting explanations for the ads they actually encountered while browsing the web. We found that users wanted explanations for around 30% of the 4,251 ads we asked them about during the study, and that subjective perceptions of how their personal data was collected and shared were highly correlated with when users wanted ad explanations. Often, users wanted these explanations to confirm or deny their own preconceptions about how their data was collected or the motives of advertisers. A key upshot of our work is that one-size-fits-all approaches to ad explanations are likely to fail at addressing people’s lived concerns about ad targeting; instead, more personalized explanations are needed.

Centralized online social networks --- e.g., Facebook, Twitter and TikTok --- help drive social connection on the Internet, but have nigh unfettered access to monitor and monetize the personal data of their users. This centralization can especially undermine the use of the social internet by minority populations, who disproportionately bear the costs of institutional surveillance. We introduce a new class of privacy-enhancing technology --- decentralized privacy overlays (DePOs) --- that helps cOSN users regain some control over their personal data by allowing them to selectively share secret content on cOSNs through decentralized content distribution networks. As a first step, we present an implementation and user evaluation of Image DePO, a proof-of-concept design probe that allows users to upload and share secret photos on Facebook through the Interplanetary File System peer-to-peer protocol. We qualitatively evaluated Image DePO in a controlled, test environment with 19 queer and Black, Indigenous, (and) Person of Color (BIPOC) participants. We found that while Image DePO could help address the institutional threats with which our participants expressed concern, interpersonal threats were the more salient concern in their decisions to share content. Accordingly, we argue that in order to see widespread use, DePOs must align protection against abstract institutional threats with protection against the more salient interpersonal threats users consider when making specific sharing decisions.

Makerspaces have complex access control requirements and are increasingly protected through digital access control mechanisms (e.g., keycards, transponders). However, it remains unclear how space administrators craft access control policies, how existing technical infrastructures support and fall short of access needs, and how these access control policies impact end-users in a makerspace. We bridge this gap through a mixed-methods, multi-stakeholder study. Specifically, we conducted 16 semi-structured interviews with makerspace administrators across the U.S. along with a survey of 48 makerspace end-users. We found four factors influenced administrators' construction of access control policies: balancing safety versus access; logistics; prior experience; and, the politics of funding. Moreover, administrators often made situational exceptions to their policies: e.g., during demand spikes, to maintain a good relationship with their staff, and if they trusted the user(s) requesting an exception. Conversely, users expressed frustration with the static nature of access control policies, wishing for negotiability and for social nuance to be factored into access decisions. The upshot is that existing mechanisms for access control in makerspaces are often inappropriately static and socially unaware.