Data donation has received attention as a more consensual means of collecting personal data for scientific inquiry and AI technology. Yet the nature of data often donated–such as harmful online messages and menstrual tracking logs–carries risk of retraumatization (the forced reliving of traumatic experience). While the well-being of data donors is considered in prior work, approaches to retraumatization remain ad hoc. We present Trauma-Informed Data Donation (TIDD): a context-specific, exploratory design framework for adapting the Trauma-Informed Approach (TIA) from the Public Health domain to data donation. TIDD was the product of a 2-year research through design process with experts on sexual violence and trauma, and observational interviews of data donors. We use a case study applying TIDD to our custom data donation platform, Ube, as an invitation for designers to consider how TIDD could be used as a malleable foundation for donation of data associated with other forms of trauma.

Large Language Models (LLMs), pre-trained on massive text corpora, exhibit remarkable human-level language understanding, reasoning, and decision-making abilities. However, they tend to memorize unwanted information, such as private or copyrighted content, raising significant privacy and legal concerns. Unlearning has emerged as a promising solution, but existing methods face a significant challenge of over-forgetting. This issue arises because they indiscriminately suppress the generation of all the tokens in forget samples, leading to a substantial loss of model utility. To overcome this challenge, we introduce the Targeted Information Forgetting (TIF) framework, which consists of (1) a flexible targeted information identifier designed to differentiate between unwanted words (UW) and general words (GW) in the forget samples, and (2) a novel Targeted Preference Optimization approach that leverages Logit Preference Loss to unlearn unwanted information associated with UW and Preservation Loss to retain general information in GW, effectively improving the unlearning process while mitigating utility degradation. Extensive experiments on the TOFU and MUSE benchmarks demonstrate that the proposed TIF framework enhances unlearning effectiveness while preserving model utility and achieving state-of-the-art results.

Minors are at risk of myriad harms online, yet online dating romance scams are seldom considered one of them. While research on romance scams in Western countries finds victims to be predominantly middle-age, it is unknown whether minors in geographic regions with cultural norms around teenage marriage are uniquely susceptible to online dating romance scams. We present an interview study with 16 victims of online dating romance scams in Iran who were minors when scammed. Our findings show that, with Western dating apps banned in Iran, scammers locate teenage victims through messaging platforms tethered to local neighborhoods, offering relief from parental pressures related to finding a marital partner and academic performance. Using threats, lies, and the exploitation of emotional attachment lacking from their families, scammers pressured minors into financial and sexual favors. This study demonstrates how local cultural context should be foregrounded in future research on, and solutions for, technology-mediated harm against minors. Content Warning: This paper discusses sexual abuse.

Generative artificial intelligence (AI) adds unprecedented scale to capabilities for self-presentation online that may diverge from one’s physical-world identity, thus potentially misinforming consent to intimate interactions, such as in online dating. Yet there is little empirical understanding of AI-generated self-presentation and (non-)disclosure to interaction partners. We present a qualitative survey of 113 online daters who used AI-generated content in their profiles or messages seen by in-person meeting partners. Findings show that generative AI is often used to fabricate attractive dating personalities through profile text and bios, with no relevance to one’s actual identity, and is seldom disclosed to meeting partners to avoid romantic rejection. Because sexual assault is defined by mis- or under-informed consent, the study positions generative AI as a potentially significant sexual assault risk factor through its use for presentation of non-physical traits that are influential to dating outcomes yet not readily identified as AI-generated upon meeting face-to-face.Content warning: this paper discusses forms of sexual violence including rape by deception.

Trauma pervades contemporary HCI research, particularly research into sexual harm. This has prompted careful consideration of whether empirical methods that require recruitment of study participants, such as survivors of sexual violence, are worth the risk of retraumatization, as well as exploration of trauma-informed alternatives to empirical methods. In this paper, we frame design fiction (a non-empirical, speculative design practice) as a trauma-informed method when deliberate structures for mitigating retraumatization of researchers are incorporated into fiction writing. Our research team applied trauma-informed design fiction to collaboratively write a speculative sex education curriculum for visualizing consent to sex called Spatial Consent. Grounded in a future where AR/VR glasses are widely adopted, Spatial Consent provides prescriptive guidance to high school students for how to visually (as opposed to vocally) exchange consent so as to mitigate unwanted sex stemming from misunderstandings of consent. The paper makes two distinct contributions: methodological guidance on when and how to use trauma-informed design fiction, and implications of Spatial Consent on HCI research into consent and sexual violence.

Technology-supported group work, and research surrounding it, has evolved considerably since scholars first conferred on the topic in 1982. Generative AI in particular has instilled swift and considerable change to how we work and who (or what) we work with, as have other emerging technologies like augmented and virtual reality (AR/VR). This issue of the Proceedings of the ACM on Human-Computer Interaction features contributions across a wide range of topics pertaining to and orbiting group work, which encapsulate both time-tested and exciting new themes in how computing technology supports, expands, and challenges the ways individuals interact and collaborate. Like previous GROUP editions, this issue features design fiction contributions as well as note-style and full-length papers. Topics featured across these articles are expansive in terms of the technologies studied, their context of use, and their impact on interaction. This issue explores misinformation on local communities, AI as a collaborative assistant, and storytelling among families, as just a few examples.

Peer Reviewed

Online dating romance scams involve coercing online daters out of their personal property. Money is almost always framed as the target of such scams, however this may be a consequence of prior work's focus on victims from Western countries and scammers from Africa. There is relatively little knowledge of online dating romance scams that occur entirely within non-Western cultures. To address this we conducted an interview study with 20 victims of online dating romance scams in Iran. Findings show that romance scams were seldom motivated by money, but rather sexual access to a woman's body through feigning intent to get married (''body scams''). Body scammers were abetted by cultural rules that justify murder and legal punishment of women for premarital sex, thus forcing them into silence. The paper concludes with calls for more expansive definitions of online dating romance scams and the foregrounding of local cultural dynamics in their perpetuation. Content Warning: This paper discusses sexual violence.

Data donation is a growing form of personal data collection that foregrounds consent and conscious participation of the data donor. There remains little guidance on who must consent to data donation, particularly when the data represents multiple people. We provide empirical perspectives on this question through in-situ observation and interviews (N=18) with online daters who chose to donate messaging interactions with potential sexual partners for sexual violence research. Findings elucidate two diverging perspectives. Participants advocating for ''unilateral consent'' argued that consent of their messaging partner is not necessary, in part, because the anticipated benefit of data donation superseded consent. Participants advocating for ''collective consent'' wanted both messaging partners to consent to its donation, citing concerns for privacy of, and personal relationships with, the other person. Findings suggest that collective consent interfaces should be incorporated in data donation platforms, even if not strictly required by legal regulation, to improve donation of multi-person data.

Despite the absence of consent being a defining quality of computer-mediated sexual harm, there is an absence of consent models that explicitly prescribe how consent to sexual activity should be asked for, given, and denied when mediated by technology. HCI literature has advocated for the adoption of affirmative consent (''yes means yes''); however, this model was created in 1991 without consideration for computers and has been historically underutilized. Through a speculative study of VR dating with 16 women and LGBTQIA+ stakeholders, we contribute archetypes of four new computer-mediated consent models for sexual activity. These include 1) visual consent through AR/VR rather than verbal dialogue, 2) agent-mediated consent where AI agents communicate consent on behalf of sexual partners, 3) a two-layer consent process called consent-to-stimulus, and 4) environmental consent where virtual environments scaffold behaviors that can(not) be consented to. We conclude by reflecting on which models could potentially supplant affirmative consent to better mitigate computer-mediated sexual violence and harassment. Content warning: This paper discusses forms of sexual violence including rape.

A majority of people experience trauma, spurring calls to incorporate trauma-informed approaches (TIA) from public health and social work into technology design. While technologies touted as trauma-informed are starting to propagate the literature, there persists a gap in knowledge around how design teams apply TIA and qualify their technology as adhering to trauma-informed principles. We address this through a 12-month development project with trauma and sexual violence experts to produce Ube, a data donation platform for collecting online dating sexual consent data to improve sexual risk detection AI. Through analysis of design documentation we retrospectively articulate a trauma-informed design process that evolved through the course of Ube’s development, comprising three elements for integrating trauma-informed principles: design goals that adapt the definition of TIA to the application domain, design activities that map to trauma-informed principles, and consequent design choices. We conclude with methodological recommendations to improve trauma-informed design processes.

On online platforms, algorithms help us build and manage our relationships. However, their invisible interventions might also pose harm to these connections. Dating platforms offer a prime example where, despite extensive research on human-inflicted harm, the potential harm from the algorithms themselves, and user strategies for mitigating them, remains largely unexplored. In our analysis of 7,043 reviews and interviews with 30 Tinder users, we unveiled how users perceive algorithmic harm as damaging self-esteem, sabotaging potential relationships, encouraging antisocial behavior, and misrepresenting or marginalizing certain identities. We introduce a new algorithmic folk theory, the "conflict of interest" theory, perceived to perpetuate these harms. This theory encapsulates users' sense of a contradiction between the dating platform's promise of finding the perfect partner (leading to discontinued use of Tinder) and its commercial interest in retaining users to increase revenue. Users suspected various algorithmic processes pursuant to this theory, such as (a) throttling profile visibility, (b) manipulating users' matches, and (c) recommending large quantities of profiles that will not lead to matches. They also described various strategies in resistance or defense of these suspected algorithmic processes, such as engaging in counter-intuitive behaviour to disrupt the unfavorable algorithmic processes or leveraging location based filtering for match variety and safety. We conclude by discussing how the perceived algorithmic harms can inform the development of new algorithmic implementations that balance both user and company interests.

Social computing platforms facilitate interpersonal harms that manifest across online and physical realms such as sexual violence between online daters and sexual grooming through social media. Risk detection AI has emerged as an approach to preventing such harms, however a myopic focus on computational performance has been criticized in HCI literature for failing to consider how users should interact with risk detection AI to stay safe. In this paper we report an interview study with woman-identifying online daters (n=20) about how they envision interacting with risk detection AI and how risk detection models can be designed pursuant to such interactions. In accordance with this goal, we engaged women in risk detection model building exercises to build their own risk detection models. Findings show that women anticipate interacting with risk detection AI to augment - not replace - their personal risk assessment strategies. They likewise designed risk detection models to amplify their subjective and admittedly biased indicators of risk. Design implications involve the notion of personalizable risk detection models, but also ethical concerns around perpetuating problematic stereotypes associated with risk.

This paper uses online dating as a context to explore futures for sexual consent technology: systems that mediate how partners exchange consent in order to prevent nonconsensual sex. Motivated by evidence that sexual consent is already mediated by computers in ways that challenge perceptions of sexual agency, we present a participatory design study in the United States with 17 women and LGBTQ+ stakeholders (demographics at disproportionate risk of sexual violence). Contrary to consent apps that are used right before sex to record irrevocable consent, participants envisioned alternative consent technology being used across online and offline interaction to normalize candid dialogue about sexual expectations and informed verbal consent throughout sex. Findings demonstrate opportunity for dating apps and associated technologies to foster voluntary adoption of affirmative consent, which has been widely advocated in public health for sexual violence prevention yet historically under-adopted by the general public. Content warning: graphic descriptions of sexual activity.

Ethics have become a core consideration in human-robot interaction (HRI) due to ample opportunity for both positive and negative impact on humans. HRI literature has expounded on ways to produce ethical social robots, especially participatory design (PD) that integrates anticipated users and other stakeholders as designers themselves to ensure their values are integrated into robot design. We draw attention to the ethics of participation in robot design, distinct from the ethics of the robot ultimately designed. We propose an approach to foregrounding ethics in PD processes through co-construction of robot PD protocols with stakeholders. We call this ”pre-PD” because it entails expanding the boundaries of PD beyond the product of design (the robot) to also include the participatory activities that enable design. Contributions of the paper include: (1) a case study of pre-PD for sexual violence mitigation robots to demonstrate feasibility of stakeholders co-constructing robot PD protocols, and (2) an actionable framework for HRI researchers to use when constructing their own PD protocols with stakeholders, informed by reflection on the case study.

The design of social matching and dating apps has changed continually through the years, marked notably by a shift to mobile devices, and yet user safety has not historically been a driver of design despite mounting evidence of sexual and other harms. This paper presents a participatory design study with women-a demographic at disproportionate risk of harm through app-use-about how mobile social matching apps could be designed to foreground their safety. Findings indicate that participants want social matching apps to augment women's abilities for self-protection, reflected in three new app roles: 1) the cloaking device, through which the social matching app helps women dynamically manage visibility to geographically nearby users, 2) the informant, through which the app helps women predict risk of harm associated with a recommended social opportunity, and 3) the guardian, through which the app monitors a user's safety during face-to-face meetings and augments their response to risk.

As social Virtual Reality (VR) grows in prevalence, new possibilities for embodied and immersive social interaction emerge, including varied forms of interpersonal harm. Yet, challenges remain regarding defining, identifying, and mitigating said harm in social VR. In this paper, we take an alternative approach to understanding and designing solutions for interpersonal harm in social VR through the lens of consent, which circumvents the lack of consensus and social norms on what should be defined as harm in social VR and reflects the embodied, immersive, and offline-world-like nature of harm in social VR. Through interviews with 39 social VR users, we offer one of the first empirical explorations on how social VR users understand consent as "boundaries," (re)purpose existing social VR features for practicing consent as "boundary setting," and envision the design of future consent mechanics in social VR to balance protection and interaction expectations to mitigate interpersonal harm as "boundary violations" in social VR. This work makes significant contributions to CSCW and HCI research by (1) uncovering how social VR users craft novel conceptualizations of consent as boundaries and harm as unwanted boundary violations, and (2) providing three foundational principles for designing future consent mechanics in social VR informed by actual social VR users.

This paper presents a participatory design study about how consent to interaction and observation of other users can be supported in social VR. We use emerging VR dating applications, colloquially called the dating metaverse, as context for study of harm-mitigative design structures in social VR given the evidence of harms that occur through dating apps and general social VR applications individually, and the harms that may occur through their convergence. Through design workshops with potential dating metaverse users in the Midwest United States (n=18) we elucidate nonconsensual experiences that should be prevented and participant-created designs for informing and exchanging consent in VR. We position consent as a valuable lens for which to design preventative solutions to harm in social VR by reframing harm as unwanted experiences that happen because of the absence of mechanics to support users in giving and denying agreement to a virtual experience before it occurs.

Use and design of dating apps has evolved in recent years to accommodate other interaction goals beyond dating, prompting some researchers to now refer to these apps as people-nearby applications (PNAs). With this expansion of use comes increased potential for misinterpretation of users' goals for meeting face-to-face, which can pose risks to user safety particularly when disparities in sexual expectations occur. We present a survey study (n=132) with users of several PNAs and with various motivations for PNA-use to understand how interaction goals are disclosed and detected. The study finds such practices to be far from consistent, with some appearing highly susceptible to misinterpretation such as purposely delaying self-presentation of interaction goals until meeting face-to-face, and implying sexual expectations through vague references to "fun." We conclude by suggesting a modified version of the "swiping" feature in PNAs to facilitate consistent and overt self-presentation of interaction goals in ways that accommodate user tendencies discovered in the study.

This study investigates how different character perspectives (a main vs. supporting character) impact story-consistent attitudes and beliefs and how that effect varies as a function of modality (mobile phone vs. VR headset) in an immersive story for preventing binge drinking. Results of a controlled experiment (N = 147) showed that participants assuming the perspective of a binge drinker (the main character) reported less negative attitudes toward binge drinking and less perceived severity of binge drinking than those assuming the perspective of a non-binge drinking friend (the supporting character). Moreover, participants who watched the story in VR headsets (vs. phones) reported greater perceived susceptibility of binge drinking. A moderated mediation effect was found, suggesting that psychological reactance significantly mediated the effect of perspective on attitudes when the immersive story was viewed in phones. This study highlights the importance of character perspective and viewer’s modality when designing immersive stories for health.

This paper reports an interview study about how consent to sexual activity is computer-mediated. The study's context of online dating is chosen due to the prevalence of sexual violence, or nonconsensual sexual activity, that is associated with dating app-use. Participants (n=19) represent a range of gender identities and sexual orientations, and predominantly used the dating app Tinder. Findings reveal two computer-mediated consent processes: consent signaling and affirmative consent. With consent signaling, users employed Tinder's interface to infer and imply agreement to sex without any explicit confirmation before making sexual advances in-person. With affirmative consent, users employed the interface to establish patterns of overt discourse around sex and consent across online and offline modalities. The paper elucidates shortcomings of both computer-mediated consent processes that leave users susceptible to sexual violence and envisions dating apps as potential sexual violence prevention solutions if deliberately designed to mediate consent exchange.

This paper presents a study of a messaging interface prototype for online dating intended to improve women's face-to-face meeting decisions, and therefore the capacity to manage the gendered risks involved with such meetings. The interface prompts users with discussion topics that are potentially more valuable for user evaluation than the impression management-motivated topics often chosen by men. These topics come in the form of first-date scenarios that messaging partners either agree or disagree on. Through a mixed-methods study utilizing speed dating events, daters used the interface to interact before meeting face-to-face. Results indicate that women's face-to-face meeting decisions improve when the interface prompts them to discuss scenarios involving agreement of opinion. Men's decisions are worsened by the same interface variant, potentially due to the displayed agreement being misinterpreted as a signal of compatibility. The study ultimately stresses that designs intended for women, and at-risk groups more broadly, must also be assessed with other user demographics-namely those that pose a risk-to identify unforeseen implications.

This paper reports on efforts to design a social matching system that instigates collaborative research across multiple fields of practice, in this instance: researchers from academia and organizations in their local geographic community. A qualitative study is presented about university researchers and the design of their profile pages for the system. Findings show that university researchers prefer profile page designs that enable them to demonstrate a willingness to adapt to non-academic partners, such as by de-emphasizing esoteric markers of expertise like scholarly publications and clarifying their resources and goals. Some also wish to circumvent potential bias by omitting information about their name, physical appearance, and academic department. However, these desired omissions raise questions about how to design for sufficient distinction between profile pages and the presentation of a unique professional identity. Implications are discussed for the design of social marching systems for collaboration.

Online dating systems are a common way to discover romantic partners. Yet there persists a gap in knowledge regarding how users of these systems determine which potential partners are worthy of in-person meetings, as well as the outcomes of these in-person meeting decisions. The objective of this dissertation is two-fold: 1) to understand how online dating system users make decisions to meet or not meet potential romantic partners in-person, and 2) to understand how online dating system designs currently support--and could better support--predictions of initial in-person attraction to potential romantic partners.

Despite the popularity of online dating systems, prior work indicates that online daters struggle to present and evaluate traits germane to attraction. However, the research community is largely unaware of successful online dating system-use strategies and how the knowledge of these strategies could inform system design. In this paper we begin to address this gap in knowledge with an interview study of online dating coaches who consider their advocated system-use strategies successful based on personal use and feedback from clients who implemented their strategies. Their advocated strategies entail keeping online evaluation intentionally minimal and persuading users to meet in-person as quickly as possible because current system designs stifle reliable evaluation of attraction-relevant traits online, notably personality. We discuss the implications of insights gathered from these strategies for online dating system design using theories relevant to romantic attraction.

Online dating systems are the most common way people meet their marriage partners online. Nevertheless, online daters struggle to evaluate personality traits of potential partners using profile pages and private messaging in these systems. Meanwhile, Multiplayer Online Games (MOGs) have emerged as a popular way young people find romantic partners for relationships in the physical world. We conducted two interview studies -- one concerning evaluation behavior in online dating systems (n=41) and the other concerning collaborative activities in MOGs (n=35). Insights from these studies reveal the weaknesses in evaluation tools native to online dating and suggest that collaborative activities could potentially address evaluation challenges in online dating. The paper concludes with a discussion of a series of design concepts for online dating in order to improve users' abilities to evaluate their potential romantic partners for in-person meetings.

Despite the importance of individuals coming together for social group-activities (e.g. pick-up volleyball, chess clubs), the process by which such groups coalesce is poorly understood. Existing theories focus on adoption and contribution rates, group types, and the formation of group norms, as opposed to the processes involved in initial group coalescence. We address this gap in the literature through an interview study examining: 1) how well people's needs for social group activity engagement are being met; 2) the challenges they face in finding and participating in, and; 3) leading interest-based group activities. Our findings highlight how people-s needs are not being addressed by current technologies. In particular, they place a heavy burden on individuals to step forward into leadership positions where the return they will receive for their efforts is often unknown, or extremely limited. We discuss the implications of our findings for the design of interest-based group coalescing technology.

Online dating systems are now widely used to search for romance and yet there is little research on how people use these systems to manage their impressions with potential romantic partners. To address this issue we conducted an interview study of 41 online dating users, revealing that-contrary to prior work-online daters largely do not want to intentionally deceive their online dating partners because they think such lies would quickly be discovered face-to-face. Nevertheless, bad first dates were a norm rather than an exception for this study's participants. In this paper we present various frustrations online daters associate with conveying and forming impressions of potential romantic partners before meeting face-to-face. We discuss the implications of these findings for the design of online dating systems.