The Sociotechnical Stack: Opportunities for Social Computing Research in Non-Consensual Intimate Media

Non-consensual intimate media (NCIM) is the unauthorized creation, sharing, or distribution of sexual content containing someone's body or likeness. This form of online abuse usually targets women, who constitute about 90% of all cases [1,33,53]. NCIM cases can be varied, ranging from sexually explicit deepfakes, to covert recordings, to sextortion (See Fig. 2). NCIM is common-studies suggest 1 in 6 adults have had their intimate content shared without consent [77,81].

NCIM is also deeply traumatic-almost all victim-survivors experience severe anxiety and depression. About half seriously consider ending their own lives, and in some cases, women and girls commit suicide to escape emotional pain and social shame. [26,33,53]. While there is widespread agreement that NCIM is harmful and some agreement that it should be regulated, there is less consensus on how to do so [22,23,28,35]. The law is evolving to recognize NCIM, but access to legal recourse is limited and cannot prevent continued re-distribution. Victim-survivors may request content removal directly from platforms, but these efforts are typically laborious, emotionally taxing, and frequently futile due to the diverse jurisdictions governing online platforms [28,66].

Effectively addressing NCIM requires both sociotechnical and computing research. Sexual abuse and patriarchal socio-historical conditions have enabled power and domination of others, typically women, via sex and intimacy.

However, new technology has weaponized this at an unprecedented scale.

This paper highlights the essential links between NCIM and its enabling technical components. Drawing on existing surveys and interviews with more than 400 victim-survivors, we identify a critical gap: technologies, including social media platforms, file formats, and content recommendation algorithms, do not merely enable but actively create NCIM at scale. To systemically analyze this, we introduce a conceptual framework called the sociotechnical stack.

In software engineering and computer science, a "tech stack" refers to a layered structure of technologies used to build systems, like the TCP/IP protocols in computing networking. We extend this concept to include social impacts, integrating the technical and social dimensions of systems. Our analysis shows how certain layers, such as a content recommendation algorithm, can unintentionally promote NCIM. The sociotechnical stack illustrates the ways our digital environment facilitates risks and harms for NCIM victim-survivors, including the non-consensual creation of sexual content via generative AI and unauthorized redistribution of privately shared intimate media. We propose a roadmap for sociotechnical research to address NCIM and alleviate risks for victim-survivors. This paper's main contributions are:

• The sociotechnical stack conceptual framework: Intended for analyzing social computing issues, the sociotechnical stack framework breaks down complex sociotechnical problems into specific technical components that contribute to social impacts and harms.

• A sociotechnical analysis of NCIM: This analysis sheds light on how various elements of the technical stack facilitate NCIM. We identify gaps in the current research on NCIM and demonstrate how adopting a sociotechnical perspective can help address these issues.

• A research agenda for NCIM: Finally, we propose important areas for further computing research on NCIM, at different layers of the sociotechnical stack.