Continuous location sharing (CLS) applications are widely used for safety and social convenience. However, these applications have privacy concerns that can be used for control and harm. To understand user concerns, we performed the largest user study of CLS application usage performed to date, with 1500 of 3000 users indicating they use CLS applications and 896 of these users completing surveys. From survey responses, we conducted 23 interviews with participants who had uncomfortable experiences. With these interviews, we perform thematic analysis grounded by sociological frameworks of power dynamics and social exchange theory. We observe that CLS application users face discomfort related to three primary categories that build on each other: (1) overstepped boundaries, (2) continued discomfort, and (3) lifestyle-impacting behaviors. With this foundational understanding, we suggest features that aim to reduce relationship imbalances that CLS applications enable. Our resulting study demonstrates that CLS applications contribute to interpersonal discomfort, highlighting the need for design changes.

Audio deepfakes represent a rising threat to trust in our daily communications. In response to this, the research community has developed a wide array of detection techniques aimed at preventing such attacks from deceiving users. Unfortunately, the creation of these defenses has generally overlooked the most important element of the system - the user themselves. As such, it is not clear whether current mechanisms augment, hinder, or simply contradict human classification of deepfakes. In this paper, we perform the first large-scale user study on deepfake detection. We recruit over 1,200 users and present them with samples from the three most widely-cited deepfake datasets. We then quantitatively compare performance and qualitatively conduct thematic analysis to motivate and understand the reasoning behind user decisions and differences from machine classifications. Our results show that users correctly classify human audio at significantly higher rates than machine learning models, and rely on linguistic features and intuition when performing classification. However, users are also regularly misled by pre-conceptions about the capabilities of generated audio (e.g., that accents and background sounds are indicative of humans). Finally, machine learning models suffer from significantly higher false positive rates, and experience false negatives that humans correctly classify when issues of quality or robotic characteristics are reported. By analyzing user behavior across multiple deepfake datasets, our study demonstrates the need to more tightly compare user and machine learning performance, and to target the latter towards areas where humans are less likely to successfully identify threats.

Reproducibility is crucial to the advancement of science; it strengthens confidence in seemingly contradictory results and expands the boundaries of known discoveries. Computer Security has the natural benefit of creating artifacts that should facilitate computational reproducibility, the ability for others to use someone else's code and data to independently recreate results, in a relatively straightforward fashion. While the Security community has recently increased its attention on reproducibility, an independent and comprehensive measurement of the current state of reproducibility has not been conducted. In this paper, we perform the first such study, targeting reproducible artifacts generated specifically by papers on machine learning security (one of the most popular areas in academic research) published in Tier 1 security conferences over the past ten years (2013-2022). We perform our measurement study of indirect and direct reproducibility over nearly 750 papers, their codebases, and datasets. Our analysis shows that there is no statistically significant difference between the availability of artifacts before and after the introduction of Artifact Evaluation Committees in Tier 1 conferences. However, based on three years of results, artifacts that pass through this process work at a higher rate than those that do not. From our collected findings, we offer data-driven suggestions for improving reproducibility in our community, including five common problems observed in our study. In so doing, we demonstrate that significant progress still needs to be made in computational reproducibility in Computer Security research.

Automatic speech recognition and voice identification systems are being deployed in a wide array of applications, from providing control mechanisms to devices lacking traditional interfaces, to the automatic transcription of conversations and authentication of users. Many of these applications have significant security and privacy considerations. We develop attacks that force mistranscription and misidentification in state of the art systems, with minimal impact on human comprehension. Processing pipelines for modern systems are comprised of signal preprocessing and feature extraction steps, whose output is fed to a machine-learned model. Prior work has focused on the models, using white-box knowledge to tailor model-specific attacks. We focus on the pipeline stages before the models, which (unlike the models) are quite similar across systems. As such, our attacks are black-box, transferable, can be tuned to require zero queries to the target, and demonstrably achieve mistranscription and misidentification rates as high as 100% by modifying only a few frames of audio. We perform a study via Amazon Mechanical Turk demonstrating that there is no statistically significant difference between human perception of regular and perturbed audio. Our findings suggest that models may learn aspects of speech that are generally not perceived by human subjects, but that are crucial for model accuracy.

Speech and speaker recognition systems are employed in a variety of applications, from personal assistants to telephony surveillance and biometric authentication. The wide deployment of these systems has been made possible by the improved accuracy in neural networks. Like other systems based on neural networks, recent research has demonstrated that speech and speaker recognition systems are vulnerable to attacks using manipulated inputs. However, as we demonstrate in this paper, the end-to-end architecture of speech and speaker systems and the nature of their inputs make attacks and defenses against them substantially different than those in the image space. We demonstrate this first by systematizing existing research in this space and providing a taxonomy through which the community can evaluate future work. We then demonstrate experimentally that attacks against these models almost universally fail to transfer. In so doing, we argue that substantial additional work is required to provide adequate mitigations in this space.