Abstract
Image-based sexual abuse (IBSA) refers to the non-consensual taking, creating, or sharing of intimate images, as well as threats to share intimate images. While considerable research has examined the nature, scope, and impacts of IBSA, comparatively little is known about perpetration. Drawing on a digital ethnography of 47 different websites, this article explores how users “do gender” through the online sharing of non-consensual intimate images. Using thematic analysis, we examine interactional dynamics that produce, reinforce, or reinvent gender norms within online digital spaces. We argue that IBSA is a homosocial practice that is embedded in ritualistic objectification and othering. These relational gender practices not only bestow social capital and sustain group cohesion, but they also normalize intimate image abuse and foster the emergence of other forms of gendered violence. This study highlights the need for more nuanced accounts of IBSA perpetration that attend to the social interactions among online users.
Keywords
Introduction
Image-based sexual abuse (IBSA) is an umbrella term encompassing the non-consensual taking, creating, or sharing of intimate images (photos or videos) – including threats to share intimate images with others (“sextortion”), pressuring, threatening, or coercing someone into sharing their intimate images (“sexting coercion”), digitally altering or manipulating images to falsely depict a person nude or engaged in sexual acts (“deepfake porn”), and the unsolicited and unwanted sharing of sexually explicit images (“cyberflashing”) (Henry et al., 2020). IBSA can have significant impacts on victim-survivors, including anxiety, depression, suicidal ideation, and other social, psychological, and physical harms, often co-occurring with other forms of abuse, such as sexual or domestic violence (e.g. McGlynn et al., 2021).
To date, much of the attention on IBSA has been focused on victimization, legal responses, and prevention interventions. In contrast, less is known about the perpetrators of IBSA and the online contexts in which abuse is enacted, legitimized, and amplified. Existing studies tend to frame IBSA perpetration in terms of individual pathology, sexual gratification, or retaliatory motives, yet such approaches often overlook the complex ways that social interactions shape, reinforce, amplify, and justify abusive behavior.
In this article, we shift the analytical focus from individual motives to collective social practices, using the concept of “doing gender” (West and Zimmerman, 1987) to examine how IBSA perpetration is embedded in interactional performances of masculinity online that not only reflect broader gender norms and power relations, but also actively shape them (DeKeseredy and Schwartz, 2016). Drawing on a digital ethnography of 47 websites that host or facilitate non-consensual intimate image-sharing, we explore the constitution of gender through the interactions among users in ways that reinforce status hierarchies, promote group belonging, and reinforce broader gender power relations. Our analysis focuses on sites used by heterosexual men, where the majority of images shared non-consensually feature women. We argue that IBSA is fundamentally a communicative practice through which users enact, consolidate, and police dominant gender norms.
This perspective builds on sociological traditions of symbolic interactionism and ethnomethodology, particularly the idea that gender is not a fixed identity but an ongoing routine accomplishment, embedded in everyday interactions. West and Zimmerman (1987) describe gender as an “achieved property of situated conduct,” maintained through accountability to normative gender expectations. We also draw on concepts such as homosociality, ritual, and “othering” as specific examples of online interactions that work to uphold group solidarity, demarcate insiders and outsiders, and regulate acceptable behavior according to proscribed norms and rules. In these online communities, members not only share abusive content but also actively participate in a set of shared practices that constitute ritualized performances of masculinity, including image tributes, degrading commentary, and gamified status systems. As we argue, these practices both reflect and reinforce structural inequalities and misogynistic norms while providing a sense of belonging and recognition among users.
Literature review
There is widespread acknowledgment that IBSA causes significant harm to victim-survivors, yet little is known about those who perpetrate this form of abuse. Previous studies have shown that IBSA is most often perpetrated by men and younger adults (Henry and Beard, 2024), those with Dark Tetrad subclinical personality traits (sadism, psychopathy, Machiavellianism, and narcissism), as well as those who have had previous experiences of either perpetration or victimization (Clancy et al., 2019, 2020; Henry et al., 2020). Research suggests that motivations underlying IBSA offending typically relate to social rewards or individual benefits, such as seeking social validation or amusement, with one study finding that motivations related to bragging and improving social status were significantly more likely to be reported by men compared with women (Clancy et al., 2019). In contrast, motivations related to revenge or punishment were less commonly reported (Clancy et al., 2019, 2020; Henry et al., 2020; Ruvalcaba and Eaton, 2020).
Few qualitative studies have explored the motivations, characteristics, and behaviors of IBSA perpetrators. In one study involving semi-structured interviews with perpetrators, Mortreux et al. (2019) found that most of the 16 Australian participants they interviewed normalized, downplayed, and minimized their behavior, blamed the victims, and overall demonstrated little remorse. In another qualitative study involving interviews with nine Danish perpetrators, Jorgensen and Demant (2021) identified two types of sharing environments: (1) acquaintance-based environments, where sharers know each other and share intimate images on social media; and (2) anonymous environments, where users trade images on the Dark Web. In anonymous environments, sexual satisfaction was identified as a central motivation, whereas in acquaintance-based settings, participants were more likely to be motivated by “social membership” reasons.
Researchers have also conducted qualitative studies to examine IBSA perpetrator activity online, exploring practices which cannot be readily captured through interview- or survey-based research. Overall, these studies demonstrate high levels of heteronormality, misogyny, and the objectification of women on different websites hosting non-consensual intimate images (Hall and Hearn, 2018; Henry and Flynn, 2019; Huber, 2023; Langlois and Slane, 2017; Semenzin and Bainotti, 2020; Uhl et al., 2018). Studies that examined “revenge porn” websites reported that motivations for posting non-consensual content often revolved around an intent to shame the victim. For instance, Hall and Hearn’s (2018) analysis of MyEx.com (which was shut down in 2018) found that men who shared non-consensual images appeared to be retaliating for supposed misdemeanors by a former partner. Other studies suggest that motivations tend to differ across platforms. For example, Henry and Flynn’s (2019) examination of 77 different websites found that sexual gratification, peer networking, and social status appeared to play a much greater role in motivating participation in abusive online behavior compared with that of revenge.
Another common theme across several studies was the categorization and objectification of women depicted in the non-consensually shared images. For example, in Langlois and Slane’s (2017: 125) study of MyEx.com, they found that the victims depicted in the images were often dehumanized; in their words, reduced to a “piece of meat” or merely “holes.” Similar findings were also noted in Uhl et al.’s (2018) analysis of seven non-consensual pornography websites where women depicted in the intimate images were often referred to by specific body parts instead of as individuals. This objectification and sense of entitlement treats women as “raw material and objects of consumption,” serving to satisfy men’s sexual desires and strengthen bonds with one another (Semenzin and Bainotti, 2020: 7).
Researchers have analyzed the culture, behaviors, and interactions among users on digital platforms, finding that it reflects a form of homosocial bonding among men who seek to represent themselves as “real men” within an online community of peers (Hall and Hearn, 2018) or in the case of an upskirting site, as homosocial bonding and craftsmanship (e.g. users providing advice to other users on camera angles) (Hall et al., 2022). In their covert study of 50 channels and groups on one site, Semenzin and Bainotti (2020) found that there was a sense of brotherhood and solidarity among men in these closed groups, with users referring to each other as “soldiers” and the action of collecting and sharing non-consensual material labeled as a “mission.” Semenzin and Bainotti (2020) found that men participated in “shitstorm activities” aimed at abusing, harassing, and stalking victims on the platform. Further studies have observed how perpetrators seek approval and acceptance from other men on various sites, particularly across community forums and imageboard websites (Henry and Flynn, 2019; Huber, 2023). Previous studies also highlight the strict adherence to rules across certain sites and closed private groups. For example, Semenzin and Bainotti (2020) observed in one group that the “only rule” was for users to send intimate images of their current or former girlfriends, with failure to comply resulting in being “kicked out” of the group.
While the research literature has significantly expanded our understanding of how IBSA is perpetrated online, there are some gaps. First, previous studies have predominantly focused on a single site alone (Hall and Hearn, 2018; Langlois and Slane, 2017; Semenzin and Bainotti, 2020) or a small number of sites (Huber, 2023; Uhl et al., 2018). By focusing on just one or two sites, there is a risk that the diversity of practices, norms, and dynamics that shape IBSA across different online communities will be overlooked. Examining a range of unique sites enables researchers to capture common practices and dynamics across different communities of practice, offering a more comprehensive and nuanced understanding of IBSA.
Second, with the exception of Henry and Flynn (2019) and Semenzin and Bainotti (2020), all previous studies have focused on “revenge porn” sites (Hall et al., 2022; Hall and Hearn, 2018; Huber, 2023; Langlois and Slane, 2017; Uhl et al., 2018). Yet, motivations for IBSA are commonly tied to peer-peer communications (e.g. seeking social validation or status), with revenge-driven motivations being less commonly reported according to survey research (Clancy et al., 2019, 2020; Henry et al., 2020; Ruvalcaba and Eaton, 2020). Moreover, dedicated “revenge porn” websites, where individuals post intimate images and identifying details to “get back” at an ex-partner or other known person, have largely disappeared on the Internet due to increased legal and regulatory pressures on both users and platforms. Given the significant shifts in the landscape in recent years, it is crucial to explore new patterns and practices across a range of sites, where revenge may not be the key driver.
Third, while there has been some attention to community dynamics in previous studies, the focus has been on either individual perpetrator motives (Hall and Hearn, 2018) or on community dynamics within a single site (Huber, 2023; Semenzin and Bainotti, 2020). Moreover, in Henry and Flynn’s (2019) study, they focused primarily on the categorization of IBSA sites, offering limited exploration of community dynamics and interactions, addressing this aspect mainly within their theoretical framework.
This current study examines a wide range of online platforms with a specific focus on community dynamics and social interactions among users. This approach is important because it shapes our understanding of how online communities encourage, enable, excuse, normalize, trivialize, and amplify IBSA, and how new trends and patterns of abuse emerge as a result. To make sense of these interactions, we draw on sociological theories on gender performativity and online communities. As outlined below, this theoretical lens offers valuable insights into how the online sharing of non-consensual intimate images both reflects and reinforces broader gender norms and power structures.
“Doing gender” in digital spaces
According to symbolic interactionist Erving Goffman (1956), gender is sociological, dramaturgical, and interactional. For Goffman, gender is performed, managed, and practiced both behind the scenes and on a metaphoric stage in front of a set of observers, who are also part of the “play.” Social actors perform their gender according to a pre-existing script which is embedded in a conventional fiction of the naturalness of gender as “one of the most fundamental divisions of society” (West and Zimmerman, 1987: 126). Goffman argued that individual actors do not voluntarily or consciously engage in the enactment of gender, but rather “social frames” (e.g. ways of talking, styles of dress) dictate the properties and parameters of their performances.
While Goffman’s microsociological analysis of gender focused on scripts, performances, and ritualized displays, ethnomethodologist Harold Garfinkel (1967) focused on gender as a “social accomplishment.” Gender performances and presentations, argued Garfinkel, align with social norms to ensure the individual is “normative,” “legitimate,” and free from the scrutiny or sanctions of others. In other words, social interactions, as well as individual self-regulatory practices (e.g. masking “incompetencies,” acquiring skills), serve to ensure that gender is being done “correctly,” whereby individuals are held accountable by others in adherence to the existing social order.
West and Zimmerman (1987) combined Goffman (1956) and Garfinkel’s (1967) insights to develop a theory of gender as both a performance and an “accomplishment.” They claimed that above all, doing gender is “interactional and, ultimately institutional” (p. 126): “. . . a person’s gender is not simply an aspect of what one is, but, more fundamentally, it is something that one does, and does recurrently, in interaction with others” (p. 140, emphasis original). Importantly for West and Zimmerman (1987), “doing gender” naturalizes the gender order between men and women and contributes to the subordination of women by men.
Digitally mediated spaces offer fertile ground for studying enactments of gender. Our study primarily draws on West and Zimmerman’s (1987) concept of gender as an embodied, relational, and interactional social accomplishment; something that individuals “do” through repeated acts that align with culturally recognizable norms of masculinity and femininity. Gender is sustained through these interactions and through accountability to broader normative gender norms, enforced both by and with others (West and Zimmerman, 1987). We supplement this conceptual framework using other key concepts as instrumental to these interactional dynamics. First, homosociality, as articulated by Eve Sedgwick (2015), is key to understanding how men form bonds with one another through the shared objectification of women. Second, we consider the role of ritual in sustaining community norms and practices, reinforcing social order, and contributing to a shared meaning. In digital environments, ritualized acts – such as “tributes,” degradation, and ranking systems – serve to naturalize gendered power hierarchies and legitimate and normalize non-consensual behaviors. Third, we draw on “othering” as a further interactional and accountability mechanism for doing gender. Typically this occurs through pornified and misogynist scripts which are grounded in objectification and dehumanization, and which enable users to frame and justify their actions.
The current study
The aim of the current study was to examine the communicative and relational exchanges among online users sharing non-consensual intimate images. We conducted a digital ethnography of a range of different websites with unique architectures, including imageboards and community forums, blogging platforms, video sharing websites, and deepfake pornography sites. Our study involved exploring, documenting, organizing, and contextualizing specific activities, cultural norms, and social dynamics within these online spaces (Hine, 2000; Murthy, 2008). Fieldwork for digital ethnographies is distinct from traditional ethnographies, due to the absence of a shared material space (or “field”) and in-person interactions. Other researchers have used this method to study dark web cryptomarkets (e.g. Barratt and Maddox, 2016), online sexual deviance (e.g. Slater, 1999), as well as IBSA (Henry and Flynn, 2019; Semenzin and Bainotti, 2020). The specifics of this project, both in terms of limitations and ethical and safety considerations, required us to alter some of the more conventional methodological and analytical approaches of digital ethnography. For example, we adopted a “lurking” model of participant observation, rather than conversing directly with “informants.”
Ethics
In early 2020, ethics approval was sought from the RMIT University Human Ethics Committee and strategies were put in place to ensure the ethical conduct of this research. In consultation with Victoria Police and the Australian Federal Police, a mandatory reporting protocol for the witnessing of child sexual abuse material and other indictable crimes was established. No reportable offenses were witnessed by the researchers during the course of data collection.
As part of the ethics protocol, we held weekly 2-hour debriefing meetings to discuss emergent findings and the impacts of exposure to explicit and, in some cases, extremely disturbing content. As an additional safeguard against vicarious traumatization, we also accessed psychological services during the project – receiving financial compensation for out-of-pocket expenses. During data collection and analysis, we chronicled emergent emotional and other responses to the subject matter in a separate document. We acknowledge our positionality as feminist, sex- and tech-positive, cisgender women, with a background in gender-based violence research, which has shaped both the design and interpretation of the study. According to Lumsden (2019), reflexive practice enables researchers to acknowledge, externalize, and critically reflect on their emotions. For instance, many of our reflections revolved around not knowing the definitive context behind the image-sharing (i.e. beyond what was described in the captions and comments). This was further complicated by the knowledge that non-consensual themes, tropes, or “kinks” can be consensually mediated, rendering it possible that some women may have posted, or consented to their partners posting, their intimate images online to invite degrading and/or sexual commentary, or to more broadly “capitalize” (e.g. financially or socially) on the apparent popularity of IBSA material on these sites. Engaging in reflexive practice enabled us to challenge embedded assumptions about victimhood, consent, and sexual autonomy, which in turn enhanced our data analysis.
It is important to note that collecting data from private, member-only groups raises ethical concerns, particularly regarding informed consent. Even data or information that is available in the public realm needs to be considered in the context of the need for consent or the waiver of this requirement. To ensure our study conformed to national and international best ethical practices, we did not collect any data from private, member-only groups, and only collected non-identifying textual data from publicly accessible online sources. We treated online users as authors rather than as research participants (Willis, 2017). Given the public nature of the websites, the type of data collected, and the absence of a reasonable expectation of privacy, we determined that informed consent was not required in line with established research practices. However, we took steps to mitigate potential harms or risks by ensuring that no personally identifying information was collected about either victims or perpetrators of IBSA. To avoid inadvertently promoting harmful sites and out of respect for victim-survivors, we do not name any websites from our study. Moreover, we did not download or take screenshots of any images. This was to ensure illegal content was not inadvertently downloaded as well as to respect the privacy and dignity of those who were depicted in the images. This was considered appropriate as we were primarily focused on text-based user interactions related to the intimate images; not the images themselves. Therefore, written descriptions of the visual content of images, where appropriate, provided sufficient context for our analysis.
As a further ethical consideration, the digital ethnography was conducted exclusively on a designated laptop, which was set up by an ICT security team to be used for the purposes of this work to ensure that there was no connection to our institutional domain or personal-professional accounts. This device was connected to the Internet via a remote pocket Wi-Fi and could only be accessed from the office of the lead researcher.
Procedure
We began the study by creating a shortlist of websites that host IBSA material for analysis (the “field”), which was based on the list of sites that Henry and Flynn (2019) examined in their study. We then identified additional sites through snowball methods; for instance, through user postings, website searches, and through an extremely detailed online directory service that provided comprehensive ratings of pornography websites (which in 2025 was shut down). Included in our sample were the following categories of sites: imageboards and community forums, blogging platforms, video sharing websites, and deepfake pornography sites, all of which were accessible through straightforward surface web searches. Key characteristics across these categories are summarized in Table 1.
Description of IBSA sites included in the study.
To be included for analysis, a site needed to meet two or more of the following criteria: (1) it was specifically marketed as a repository of IBSA material (allowing for semantic variation); (2) it allowed user-generated content and/or direct uploads to the site; (3) it allowed users to comment on, or provide context around, an image and/or disclose identifying information about those depicted in the image; and/or (4) it contained pages that displayed (or potentially displayed) non-consensual intimate imagery. We defined “intimate images” to include nude, semi-nude, or sexual photos or videos.
Fieldwork occurred across five phases. First, the context and background of each online community was examined (e.g. site branding, content genres, community aims, and values) to determine eligibility for inclusion. Second, eligible sites were then analyzed in terms of the images being posted by users and the context provided for sharing (e.g. captions, album names/categories, forum categories, thread-specific objectives). Third, user interactions with one another through text-based affordances (e.g. comments, likes, and other platform features) were observed, captured, and exported/inputted as data. Fourth, community management protocols were recorded. Finally, platform architecture, affordances, policies, and terms of use were captured.
In total, we conducted observations of 47 different sites. With the exception of three sites that involved the non-consensual sharing of intimate images of gay men, the content across the dataset predominantly featured women and was shared by primarily heterosexual male users. We stopped collecting data when there were no new patterns of abuse perpetration or community interactions being observed. Relevant data were directly inputted into a spreadsheet, whereas large bodies of text (e.g. website terms and conditions) were copied and pasted into a notepad document and exported from the digital ethnography laptop using a USB. In addition, given the prohibition against image downloads, we used text to describe the general content of the images. Notably, these descriptions served only to provide context for our analysis of the types of user interactions (e.g. comments and captions) elicited by the image. The images themselves were not considered primary data.
Data coding and analysis
Thematic analysis (Braun and Clarke, 2006) was used to inductively and deductively code and analyze the data. Once data collection was complete, the first and second author, and another researcher, created a coding framework to facilitate coding the data in NVivo. Independently, each researcher read through the fieldnotes for a set of 10 websites from our research sample and inductively came up with different codes, before meeting to discuss similarities and disparities. A first draft of the coding framework was devised, which was then developed and tested against another set of 10 sites. A further meeting was convened to finalize the coding framework and to assign each three researchers an equal set of sites to code in NVivo.
The final codebook included the following codes and sub-codes: website architectures (site branding; affordances; advertising); platform policies (rules/prohibitions; liability disclaimers; transparency and consent statements); content moderation and reporting (moderation policies; format of reporting; barriers to reporting); community (user level) (expression of community; non-consensual album or thread names); abuse typology (retribution/revenge; deepfakes/cheapfakes; image requests; identification of victims; impersonation; sexual assault images; “cum tributes” and other semen and penis stuff; “hot wifing”; sexual gratification; abuse, body shaming, and degradation; and other); prejudice (racism, homophobia, transphobia, and ableism); and language (justification for abuse/victim-blaming; specific “lingo”). For data analysis and write-up, we condensed the list of codes to inductively and deductively generate a set of overarching themes.
Regarding data analysis, several researchers have used content analysis for studying different online communities, including online illicit communities sharing in child abuse material (e.g. O’Halloran and Quayle, 2010), incel groups (men who are involuntarily celibate) (e.g. O’Malley and Helm, 2023), and non-consensual nude or sexual image sharing of adults (e.g. Hall and Hearn, 2018; Uhl et al., 2018). While content analysis is appropriate for online communications that are predominantly text-based, we chose a combination of qualitative content analysis with a “virtual participant observation” method, using a digital ethnographic approach to observe platform architectures, as well as the cultural dynamics and interactions among users. In addition, the scope of the study, with regards to its objectives and research questions, required us to incorporate content analysis of both images and text (e.g. captions or user comments) with the text serving as both a discourse and a phenomenon in itself (i.e. participation in sharing or viewing abuse material within a community intended for that purpose).
Findings and discussion
Through both deductive and inductive thematic analysis, we identified three overarching (and overlapping) themes – homosociality, ritual, and othering – which we argue are examples of the interactional constituents of “doing gender” specifically in relation to the non-consensual sharing of intimate images within online communities. In the discussion of our findings below, we provide a brief explanation of these three theoretical constructs, drawing from existing sociological theories, and use these constructs as a lens for interpreting the data and contextualizing our key findings.
Homosociality: “Go ahead, she isn’t looking”
The non-consensual taking, creating, or sharing of intimate images is a means of doing gender, which involves consolidating, policing, and validating certain gender norms. To explore the social bonds that users cultivate in these online communities, we turn to the first theme of “homosociality.” Sedgwick (2015) referred to “male homosocial desire” as “a kind of oxymoron” to describe the social bonds between people of the same sex but as both distinct to, and as an analogy to, “homosexuality.” Sedgwick’s use of the word “desire” brings the homosocial into the “orbit of desire” as potentially erotic (e.g. “homoerotic”), as a social force or “glue” “even when its manifestation is hostility or hatred or something less emotively charged.” Sedgwick suggested that for men (compared with women) the continuum between homosocial and homosexual is “criss-crossed with deep discontinuities” (p. 2).
The solidarity among cis-gendered men and the discourses they promote to uphold and perpetuate gender inequalities and the gender order is, as sociologist Michael Flood (2008: 4) notes, “often accomplished through the exclusion of women and an ideological emphasis on men’s difference from and superiority to them.” Peer relationships have a strong influence on men’s sexual activity, with sexual achievements often communicated through “sexual storytelling” to other men serving as a means for proving prowess and thereby achieving status (Flood, 2008). These social interactions are often characterized as part of a “lad culture,” which refers to a “group mentality articulated through activities such as sport and heavy alcohol consumption, and characterized by sexist and homophobic ‘banter’” (Phipps and Young, 2013: 28). In our study, we identified three examples of homosociality: hot-wifing, soliciting commentary from other users, and tributes.
Hot-wifing
Across many of the online communities we examined, homosociality was exemplified by the prevalence of a voyeuristic sub-genre referred to as “hot-wifing,” “slut-wifing,” or “gf [girlfriend]-sharing,” with one site’s tagline being: “Go ahead, she isn’t looking.” The phenomenon of “hotwifing” refers to the practice of a man “sharing” his wife (or woman partner) to obtain status and sexual gratification from watching, participating in, or hearing about the pleasure other men derive from sexually dominating her. In online environments, this usually involves non-consensually posting sexual images of intimate partners, crushes, or other women (e.g. images taken from public beaches, changing rooms etc.) among the online group as a form of “offering,” often with captions like: “my wife has no idea I share her,” “enjoy my wife,” “wife swap,” and “wife to share.” Posts of this nature were typically aimed at soliciting sexual commentary and enabling the user to capitalize on the status afforded to him by her perceived “hotness,” but also owing to the taboo nature of the images. Another common practice we observed was that of “challenges” (e.g. “The Challenge Board”) where users are encouraged to set challenges for one another, such as taking surreptitious intimate images of one’s wife or girlfriend in public.
As a pertinent example of sharing non-consensual intimate images and building community through this and other practices, on one site a user shared intimate images of his girlfriend’s daughter and posted comments about his sexual attraction to her, which received more than 170 comments and several hundred-thousand views. In the user’s original post, he discussed the sexual attraction he felt toward the daughter after hearing she had posted nudes on a private account which led to him searching for the images on his girlfriend’s phone. The other users expressed their approval of his actions and feelings toward the daughter, as well as jealousy of his situation. They also encouraged him to share images of the young woman’s mother and to “blackmail them both for nudes.” There were also repeated mentions suggesting he hide cameras in the bathroom and then share “hidden cam” images among the group, as well as incitements to drug and rape her, steal her underwear, spike her drinks with his semen and then film her (unwittingly) drinking it and share the video with the group. When the original poster stopped giving updates to the group, some users took it upon themselves to create deepfakes of the daughter and recruit the “help of the best nerds” to find more photos of her.
Soliciting commentary
Another distinct interactional homosocial practice we observed was that of soliciting commentary from other users about intimate images. This typically involves requesting graphic descriptions of what other users would do to her sexually – for instance, “what would you do to my wife?” or “I am looking for a guy who would worship my wife’s curves and I’ll do the same of his wife.” During such exchanges, community members appeared to earn capital from having their “offering” to the group accepted and enjoyed by other members. Image captions often contained a call-to-action, for example, “describe how you would rape this bitch,” or a question, such as: “how can I convince my wife to let me share her with other men?” This sometimes co-occurred with doxing (i.e. revealing personally identifiable information about the person depicted in the image). Posts of this nature received high levels of member engagement through comments and likes.
Other interactive practices included users being requested to rate physical appearance and express sexual interest in the women depicted in the images; hunt down further images of women on the Internet; and digitally alter images using photoshopping, deepfake, or nudify tools – particularly of clothed or non-intimate images of women (e.g. a user sharing a woman’s bridal photos and requesting the images to be “nudeshopped”). Sometimes there were more specific trading practices, like users sharing their email addresses and offering to share intimate images of their partners in exchange for images of someone else’s partner. For example, one user wanted to organize an ongoing swap arrangement, whereby the image exchanges would be regular, and the two users could “occasionally vid chat, jerk off, cum tribute our wives pics or her things.” These highly interactive, homoerotic community practices encourage other users to make requests and return favors. For instance, one site boasts: “We’re still striving to be the best community we can be and we need YOU to make it happen. Post your hottest cuckold and interracial moments and enjoy the company of likeminded people!”
Tributes
A third striking example of homosociality is the practice of sharing “tributes.” A “tribute” is an act, statement, or gift with the aim of expressing respect, admiration, or gratitude to a thing or person. There were two forms of tributes that we identified in our study. The first is the “cock tribute,” which refers to either an image of a man masturbating to the image of someone else’s partner (e.g. either holding his penis in front of a printed-out photo or in front of a laptop or tablet screen) or an image where a penis is superimposed. The second example is the “cum tribute,” when a man ejaculates onto a printed or digital image, such as on the screen of their phone or computer, and then re-uploads it for the “pleasure” of the other users. Tributes were one of the most common forms of exchange witnessed during our ethnographic observations on most of the sites we examined, with men on one site requesting other users to “please cock her” or “cover this slut in cum.” In our field notes, we observed that “ejaculation is some kind of embodiment or manifestation of men’s dominance over women in a ritualistic fashion” and “tributes are the most consistent trend to be requested among/by users across these platforms, to reinforce community values, [and] demonstrate and test trust.”
Ritual: “everyone is expected to contribute”
Rituals are an important interactional mechanism for doing gender, and doing gender in a homosocial way. Although ritual is a contested concept, anthropological and sociological perspectives essentially treat ritual as repeated and collective actions, events, or performances. Sociologist Émile Durkheim argued that collective rituals are an important means for maintaining social cohesion – symbols that people gather around to engage in shared beliefs, sentiments, and emotions (D’Orsi and Dei, 2018). To distinguish a ritual from a habit (like brushing one’s teeth), three characteristics can be identified: first, a public dimension; second, a “sacred” (beyond the ordinary/mundane/everyday) component; and third, the emotional involvement of the social actors (D’Orsi and Dei, 2018). For Durkheim, rituals help to build systems of belief and a sense of belonging. Turner (1957) developed these ideas further, focusing on the dynamic and performative aspects of ritual. Turner argued that rituals help resolve social dramas and social conflict, to allow people to become aware of their condition and give meaning to their existence.
Like any other group, online communities have unique rituals embedded in norms, actions, rules, and language that are shared, accepted, and sometime contested and subverted by the group. These rituals normalize certain practices and allow users to engage in behaviors or express perspectives that “they may otherwise withhold due to a fear of social sanction from peers or social rejection” (Helm et al., 2024: 27). As noted by Oksanen et al. (2024), Human beings are inherently social and seek companionship and social engagement . . . Online communication responds to this social need of belonging. Online communities emerge and thrive in digital spaces, comprised of members who engage in active communication in a shared topic or interest area. (p. 1)
High identification with other group members, interacting with likeminded others, and information bias characterize online “identity bubbles” or “echo chambers” (Oksanen et al., 2024). Below we discuss how technological affordances, rules, and language enable ritualized behaviors as a means for doing gender, which in turn facilitate, encourage, and normalize gendered abuse, often instigating new forms of abuse beyond the non-consensual sharing of intimate images.
Technological affordances
Technological affordances create the conditions and architectures for ritualistic practices. Affordances are constantly evolving, and both enable and shape certain modalities of communication (Poell et al., 2018). Affordances include things like algorithms, news feeds, subscription services, private messaging, “like” buttons, upvotes and downvotes, upload and download functions, live streaming, hashtags, filters, screen sharing, channels, threads, and search functions. Affordances can also include anonymity, visibility, lack of guardianship, poor regulation, content moderation, and platform policies. For example, in our study, anonymity enables users to construct gendered narratives of themselves through the sharing of non-consensual intimate images of their women partners. The covert nature of IBSA within these online communities enables members to employ a “what they don’t know can’t hurt them” mentality to downplay the harm (e.g. “my wife has no idea I share her”) and encourage participation (e.g. “go ahead, she isn’t looking”). We also witnessed that when users confessed deviant proclivities in these anonymous contexts, they were often met with understanding and acceptance from their fellow group members. These and other community-building customs function as vehicles of normalization. For instance, on the thread mentioned above, where the user confessed his sexual desire for his girlfriend’s daughter, initially permissive replies from other users quickly translated into active encouragement, with one user posting: “[do] everything you can do. If possible, you should even go as fast [sic] as to drug and rape her!” User exchanges within the comments sections is clear evidence of how members collectively contribute to and foster enabling conditions for IBSA to take place. Yet significantly (in the case of the post described above), it is also evident that these communities come together to ritualize, gamify, and embolden acts of sexual and gender-based violence in broader offline contexts as well.
On the theme of gamification, site affordances enable ritualistic practices and in-group bonding in ways that mirror game-like mechanics. For instance, on one deepfake site, a Community page lists the “Top Active Members” which – akin to a game leaderboard – provides information about how long the users have been a member for and how many videos they have posted. One user had posted 6,262 videos and had the highest rating (i.e. rewards) for “Content Popularity” and the “Most Viewed.” Members as well as visitors to the site can upvote or downvote videos, or post comments. Members can add videos to their favorites list, mark videos to watch later, add videos to a new playlist, send messages, add friends, and chat with other users via the chat room. The site also showcases the newest members.
Similarly, many sites mirror or mimic those of social media, either in their appearance (e.g. one site that looked almost identical to Facebook) and more commonly in the types of affordances and practices that are associated with social media and apps. This arguably has the effect of legitimation and normalization, including the illusion of legality. For instance, on one deepfake site, creators can control who they share videos with and who can download them, and users can stay anonymous. Registered users can “enjoy unrestricted access to guides, support and tools.” Content creators can earn money from their downloads through having their own donation button where they set a price (in tokens) for downloading their videos, and tokens can be redeemed for payout through PayPal. Members can also benefit from “extra features” such as subscribing to creators, saving playlists, and messaging other users. To become a “verified creator” the user must first upload at least one deepfake video that they themselves made. Finally, users can commission a paid video for customized deepfakes – all done through private messaging.
Rules
Specific internal logics, rules, and self-governance practices within online communities underpin the rituals of reciprocity and exchange of capital, and the practices of not only sharing non-consensual intimate images, but also engaging in other abusive behaviors (tributes, abusive commentary, etc.). Implicit and explicit rules and norms provide the overarching framework for social interactions within online communities. For instance, on one site, the following instruction is provided: “Share all of your HOTTEST Flat Chested Cuties A-B Cup only! C cup and above will be deleted. No cocks, jizz, GIFS or captions. Multiple pics of the same girl will be frowned upon.” Another thread on the same site, entitled: “Daughters for daddies,” instructs users: “This group is for daughters that do their daddies, and daddies that do their daughters. 18+ ONLY! No scat.”
In addition to stipulated rules, social “etiquette” norms also shape user interactions and dynamics. For instance, one thread we observed was dedicated to “user requests.” Once a user’s request was fulfilled, they were instructed to “Thank the user that filled your request” and to “see if there is anyone that YOU can help out.” This highlights a noteworthy dialectic, in that despite being engaged in objectively toxic behaviors (i.e. harmful to those outside of the group), these sites often exhibited high levels of internal civility toward other members of the group (Trott et al., 2022). Another example of civility was seen on deepfake websites, whereby the technical creation of deepfake pornography was often regarded with a sense of reverence, with users leaving comments such as “this is a good fake” and moderators reminding site visitors: “Dom’t [sic] forget to like the videos and subscribe to the creators you love.” This sense of comradery and craftsmanship was seen across several sites we examined, culminating in a form of skill-sharing (e.g. how to take screenshots on Instagram) that was seen to strengthen community and normalize the dissemination of non-consensual imagery.
Language
The rituals of online communities are also characterized by specific, customized language. This is what Wetherell and Potter (1988) called “interpretive repertoire,” which refers to “internally consistent, bounded language” that people use in their social interactions with others. For instance, in our digital ethnography, examples of specific language include terms such as: “dumps” (uploaded images), “wins” (non-consensual images), “sauce” (requests for source information such as identifiable information), and “x-rays” (or “nude-shopping”) (using AI to remove clothing). Similarly, in their study of the “manosphere” (a collection of online men’s support communities predicated on anti-feminism and misogyny), Marwick and Caplan (2018) observed how community members use a specialized language to sustain the gendered social order. They note that language serves discursive and ideological functions to foster a sense of community. These online communities also heavily depend on recurring visual codes of feminine vulnerability, submission, and sexual availability. This not only marks women as appropriate targets for ritualized abuse, but the repetition of these acts contributes to what Goffman (1979) called the “hyper-ritualization” of gender display, in which visual and discursive tropes serve to reinforce gender norms and expectations.
Othering: “To Show them that they are nothing but objects”
The third and final theme we identified is “othering.” “Othered” gendered bodies are essential “objects” for “doing gender” within networked communities. This is achieved through the two interrelated processes of objectification and dehumanization.
Objectification
Objectification refers to treating or regarding a person as an object and not as a human being; “treating as a thing, something that is really not a thing” (Nussbaum, 1999, emphasis original). According to Nussbaum (1999), there are seven dimensions to objectification: instrumentality (the object is treated as a tool for self-interest); denial of autonomy (the object is treated as lacking self-determination); inertness (the object is treated as lacking in agency); fungibility (the object is treated as interchangeable with other object types); violability (the object is treated as deserving of objectification); ownership (the object is treated as something that can be owned); and denial of subjectivity (the object is treated as someone whose feelings and experiences are not important).
Objectification, we noted, was mediated through the interactions among users in the online spaces we investigated. For example, there were entire threads dedicated to sharing graphic images depicting men using women as household objects, such as footstalls, rubbish receptacles, and toilets; to sharing images “. . . of the cunts you wanna humiliate, rape and degrade. To Show them that they are nothing but objects.” Other submissions included women in cages, chained up in the kitchen, or packed into a see-through plastic case; naked women acting as “furniture” (e.g. a coffee table or ottoman) or women chained in bedrooms or cleaning the kitchen naked with a collar or chain. On one site there was a “rape group” where victims are dehumanized, including being referred to as “it.” On these sites, women are treated as objects of exchanges for social capital and as a means of communicating dominant gender norms.
Dehumanization
On some sites we examined, women were referred to as “rape meat” and otherwise regarded as animals. For instance, one thread entitled “Degrade this Pig” (which has 64.1 K members) solicits degrading comments from users about images of women. It is stipulated in the rules that the thread is for: “fat ugly FEMALE pigs to display themselves & receive the degrading/humiliation they crave & deserve, Misogyny, rape fantasies, race play, body shaming, degrading, humiliation & tasks/punishment.” It is reasonably likely that this thread includes images that are consensually posted by the women depicted within them. However, given that instances of IBSA can – and have been known to – involve an individual impersonating someone else using their intimate images (e.g. on a website or dating app), it remains likely that some, if not many or most, of the original posters are not the women depicted in the images. Supporting this rationale was the finding that non-consensual degradation threads were common across our sample. For instance, on a different site, there was a similar thread called “Degrade and Abuse” (containing nearly 400,000 posts), which involved users non-consensually posting images of women (clothed or nude) and asking other users to respond to two questions: (1) name “one degrading sex act you would do to the slut”; and (2) “What would your name for the slut be for the time you used her?” Other sites and threads aimed at non-consensually sharing intimate images of women and degrading them involved practices of body shaming, with one thread entitled “Cute faces but some say ugly privates.”
We also investigated sites where users engage in the sharing and viewing of sexual assault or “rape porn” images. We observed that there were some common characteristics: the perpetrators’ faces and genitals were typically blurred while the victim’s face and body were not; the video was set in a non-descript and dimly-lit location, such as a garage or basement; there were often more than two perpetrators and another individual filming; the assailants also physically harmed or degraded the victim, including covering her mouth with their hands, slapping her face and/or genitals, and rubbing semen onto her face; and the victims were dirty, disheveled, and appeared to be drugged or in a state of distress. The videos were also continuously shot, with no change in angles. These disturbing examples demonstrate the complexity of offending where a physical sexual assault has been perpetrated alongside the filming of the assault and the sharing of the footage.
It is also important to note that women are not the only group othered, objectified, and dehumanized in these online communities. We also noted strong themes of racism, ableism, homophobia, and transphobia. For instance, heterosexual men’s disavowal and repudiation of homosexual desire further serves to disrupt the continuity between the homosocial and the homosexual (Sedgwick, 2015). Within these deviant online communities, homosocial boundaries assist in the consolidation and “policing” of heteronormativity by users in their social interactions with each other. Homosexuality was considered the “ultimate transgression” within most of these communities, with homophobic slurs often being used to police and call out members who violated internal community rules. There was also a notable fetishization of Black men, which was observed in users’ requests for “BBC edits” (Big Black Cocks), featuring women “taking” BBC in various sexual and explicit ways.
In summary, the exchange and trade of non-consensual nude or sexual imagery online is a means through which to display or enact gender (DeKeseredy and Schwartz, 2016; Gjika, 2023), and in doing so remain accountable to social prescriptions of gender (West and Zimmerman, 1987). Users not only share non-consensual intimate images, but also give accounts of their gendered pursuits and exploits, engaging in highly degrading, objectifying, and misogynistic discourse with one another. These homosocial, ritualizing, and othering interactions normalize and reinforce the gender order, as users are held accountable to each other. It is important to note, however, that our analysis focused on sites with predominantly heterosexual male users and content involving women. We did not conduct comparative analysis of the smaller subset of sites featuring non-consensual images of gay men, and nor did we examine kink-focused sites in detail, which limits the generalizability of our findings somewhat across the broader spectrum of IBSA practices.
When examining social interactions among users within online communities, it is important to also investigate what is driving these behaviors. As noted by Helm et al. (2024), “Online communities may be particularly appealing to individuals who feel outcast or isolated in their offline lives, as communities can provide vulnerable individuals with support or validation” (p. 28). These drivers may include “precarious manhood, gender role strain, and distress . . .” and “deep-seeded feelings of injustice” (O’Malley and Helm, 2023: 1029). One limitation of our study was that we were not able to investigate the individual-level drivers of these behaviors, particularly those situated at the personal and psychological level of the ecological model of gender-based violence. Further research should explore these factors in greater depth through methods such as quantitative surveys and qualitative interviews.
Our study was also limited to content that was publicly available. We did not conduct any observations in closed, private groups due to ethical considerations. Nonetheless, we examined a large number of online communities where non-consensual intimate images are being shared, and as such, were able to examine the subculture “writ large” (O’Malley and Helm, 2023). Thus, our observations are somewhat generalizable, and point to the importance of social interactions underpinning these behaviors – interactions that prevention efforts should seek to squarely address.
Conclusion
This study sought to investigate the community dynamics underpinning the non-consensual sharing of intimate images online. Rather than isolated acts of deviance, IBSA emerges as a digitally mediated social practice, shaped through shared rituals, homosocial bonding, and community norms that valorize dominance, control, and the degradation of women and other marginalized groups. Users embody, perform, and act out their gender in these online spaces, and are held accountable by other users through group norms and rules, both written and unwritten. Mechanisms of gender performance and accountability include homosociality, ritual, and othering. Much like narratives of sexual conquest and exploits told in face-to-face interactions among men, intimate images serve as the “semiotic resources” (Van Doorn, 2010) for doing gender and building community. Through technological affordances, like anonymity, gamification, and ranking systems, these behaviors are implicitly and explicitly normalized, encouraged, and rewarded.
We also found that intimate image abuse is not only normalized and amplified in these online spaces, but that new forms of abuse are materializing. These include related – but arguably distinct – practices, such as rating women’s physical appearance, speculating about imagined sexual encounters (e.g. “how long would she last in a gangbang?”), offering advice about online and in-person abuse, digitally manipulating images using deepfake technology to generate new images, ejaculation practices (e.g. “cum shots”), or superimposing penises onto non-sexualized images.
This study reframes IBSA not merely as an act of individual violation, but as a collectively enacted and normalized communicative practice of doing gender. These insights underscore the need for further research that connects interactional performances to the wider political economies of digital culture, including the influence of pornography, platform capitalism, and global misogynistic networks.
Understanding interactional dynamics is essential for developing more effective strategies for prevention. While it is well established that primary prevention more broadly must address structural gender inequalities, such as pay inequities and discrimination, it is equally important to examine the everyday practices and interactions that sustain the gender order (Deutsch, 2007). If, as Judith Butler (1988) argues, gender performativity is “a stylized repetition of acts through time” and “not a seemingly seamless identity” (p. 520), then transformation lies in subverting those acts—“in the possibility of a different sort of repeating” (p. 520). In addition to fostering cultural change around gender, prevention efforts must address intersecting systems of inequality and growing patterns of social disconnection (including those linked to poor mental health). Prevention must also focus on mobilizing bystander intervention to disrupt the distributive flow of non-consensual imagery, and promoting digital and porn literacy, positive sexuality, harm reduction, and affirmative consent.
Footnotes
Acknowledgements
The authors gratefully acknowledge funding from the Australian Research Council. They also thank Alice Witt and Stefani Vasil for their valuable assistance, and the anonymous reviewers and editors for their constructive feedback. The authors further wish to acknowledge the contributions and support of the national and international advisory board members: Jane Bailey, Karen Bentley, Bronwyn Carlson, Asia Eaton, Mary Anne Franks, Mia Garlick, Nicola Gavey, Michelle Gonzalez, Bridget Harris, Kara Hinesley, Emma Holten, Michelle Lewis, Clare McGlynn, Elena Michael, Sophie Mortimer, Karuna Nain, Michael Salter, Zahra Stardust, Nic Suzor, and Samantha Yorke.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by an Australian Research Council (ARC) Future Fellowship grant (FT20010064).
