The capability of GenAI-based chatbots, such as ChatGPT and Gemini, has expanded quickly in recent years, turning them into GenAI Chatbot Ecosystems. Yet, users’ understanding of how such ecosystems work remains unknown. In this paper, we investigate users’ mental models of how GenAI Chatbot Ecosystems work. This is an important question because users’ mental models guide their behaviors, including making decisions that impact their privacy. Through 21 semi-structured interviews, we uncovered users’ four mental models towards first-party (e.g., Google Gemini) and third-party (e.g., ChatGPT) GenAI Chatbot Ecosystems. These mental models centered around the role of the chatbot in the entire ecosystem.We further found that participants held a more consistent and simpler mental model towards third-party ecosystems than the first-party ones, resulting in higher trust and fewer concerns towards the thirdparty ecosystems. We discuss the design and policy implications based on our results.

Smart devices--including speakers, cameras, TVs, and sensors--are increasingly used to enhance comfort, efficiency, security, and entertainment in and outside the home. These devices collect significant audio, video, and environmental data, and that data may be shared with device owners, manufacturers, and/or third-parties. Social computing researchers have devoted significant attention to the privacy risks posed by these devices, although the focus has largely been on primary users who have access and control of devices and data. More recently, researchers have begun examining the privacy risks to incidental users --those who do not own or manage devices but still have their data collected due to their presence in a smart environment. In this paper, we present findings from 25 interviews with US-based adults who have encountered and/or used smart devices as incidental users in rental properties, as part of their work, at friends' and family's houses, and in shared living environments. We explore their reactions to these devices across different contexts, as well as their expectations for how devices should be disclosed and managed by primary users. We found a prevailing sense of resignation across our participants, as well behavioral and environmental modifications in response to smart devices. At the same time, participants were disinterested in gaining access and control to other people's devices, and they were hesitant or uncomfortable with the idea of initiating communication with device owners regarding their concern. Many wanted clearer and more prominent disclosures about device and data management that shifted responsibility from the incidental user to the device owner. Based on our findings, we argue for researchers to focus more closely on sociotechnical approaches to help incidental users navigate privacy risks in smart environments. Such approaches must account for interpersonal factors like trust, power dynamics, and social norms in addition to the many technical solutions that have been introduced in recent years.

In most American cities commuters on public transit have disproportionately lower incomes than commuters who use automobiles. Given the proven link between geographic and economic mobility, it is critical to offer quality public transit to improve access to jobs, health care and education opportunities. Departments of Transportation (DOTs) routinely measure public transit performance and quality perceptions to assess the need for improvements in the transit systems. Nevertheless, the performance metrics used fail to capture the experiences of low-income individuals who often endure complex, lengthy trips, requiring several modes or transfers. We propose BALTO, a novel toolkit to characterize transit system performance and passenger’s quality perceptions across all types of passengers and trips. We are designing the BALTO toolkit in collaboration with public housing residents from the Housing Authority of Baltimore City (HABC) and together with two local transit advocacy groups and the departments of transportation in Baltimore and in the state of Maryland.

Fueled by Internet-of-Things technologies and spanning a wide range of sensors, speakers, and cameras, smart homes promise to make our lives easier and automate routine tasks. From speakers to security cameras, smart home devices (SHDs) answer our questions, monitor our home environment, and conserve energy. They also collect significant data, ranging from on/off commands to audio and video data, and they do this in some of our most private spaces. In this paper, we explore the privacy risks associated with SHDs by focusing on privacy-conscious smart home power users--those who spend significant time and money to research, install, and integrate devices throughout their homes and engage in advanced device and network management strategies to mitigate privacy concerns. Drawing on data from 10 focus groups with 32 privacy-conscious power users, we identify the key privacy risks they perceive from this technology, as well as how they mitigate those risks through increasingly complex strategies. Our findings reveal that navigating the technical landscape that makes up the smart home environment--including what data is collected, what options are available for managing or restricting data flows, and who has access to data collected by SHDs--is complex and often confusing, even for people who spend significant time researching devices and integration options. We use these findings to argue for further development of tools that are transparent, easy to use, and aligned with the privacy needs of a diverse userbase.