---

To find the workshop room, please follow the signs from the registration desk or check out the map.

Event	Time (CET)
Welcome and introduction	9:00–9:30am
Keynote address by Dr. Eleonora Esposito: The Image Is Synthetic, the Abuse Is Real	9:30–10:30am
Break	10:30–10:40am
Lightning presentations	10:40–11:15am
Break	11:15–11:25am
Small-Group discussion	11:25–11:55am
Reflection and wrap-up	11:55am–12:00pm

---

Keynote Speaker

Speaker Biography

Dr Eleonora Esposito is a Research Affiliate at the Institute for Culture and Society, University of Navarra, Spain.

She holds a European PhD in Critical Discourse Studies and was a Marie Skłodowska-Curie Individual Fellow, leading the WONT-HATE project on online violence against women in politics.

Her work lies at the intersection of gender studies, digital media, platform regulation, and critical discourse analysis, with particular attention to online abuse against women in public life, feminist digital activism, and the social and regulatory implications of digital technology.

Eleonora has previously worked as Project Manager at the European Institute for Gender Equality and currently works as a Policy Analyst and Team Leader for gender-based violence as a systemic risk at the European Commission. Her academic and policy experience informs a research agenda centered on survivor-focused approaches, platform accountability, and the governance of technology-facilitated harms.

Talk Title

The Image Is Synthetic, the Abuse Is Real: Generative AI-Facilitated Image Abuse as Technology-Facilitated Gender-Based Violence (TF-GBV)

Abstract

Generative AI has not created image-based sexual abuse, but it has profoundly transformed its scale, accessibility, and modes of production. Non-consensual intimate imagery and child sexual abuse material predate diffusion models, and the scholarship examining these practices as forms of technology-facilitated gender-based violence, TF-GBV, provides an established body of knowledge on how such harms operate, how they affect survivors, and why removal-based responses alone are insufficient.

This keynote brings that body of research into contemporary debates on AI-facilitated image-based abuse. It argues that the way harm is conceptualised directly shapes the range of solutions considered appropriate, and that treating AIG-NCII and AIG-CSAM primarily as novel technical problems risks reproducing interventions that survivor-centered research has already shown to be inadequate.

Drawing on victim-centered scholarship and the evolving regulatory landscape at EU and international levels, the talk distinguishes the specific transformations introduced by generative AI from the continuities that remain central to these forms of abuse. These include gendered power relations, sexual humiliation, coercion, reputational harm, silencing, institutional failure, and the persistent displacement of the burden of response onto victims.

The keynote examines the legal, technical, and institutional frameworks developed to address these harms and argues that effective prevention and remedy require attention to the full chain of abuse. It concludes by outlining a cross-disciplinary research and regulatory agenda spanning AI safety, technology policy, law, platform governance, human-computer interaction, and survivor-centered research.

---

Lightning Talks

Critical Considerations Around the Measurement of Deepfake Abuse and the Design of Psychoeducational Interventions

John Twomey (Presenter, Department of Psychology, University of Limerick) & Gillian Murphy (School of Applied Psychology, University College Cork)

Deepfake abuse consists of the creation of non-consensual intimate imagery through the use of AI to generate or modify existing videos. Drawing from a number of ongoing projects we are conducting in this area, this position paper aims to highlight three critical concerns to inform the future of deepfake abuse research. Firstly, the existing measurement of deepfake abuse, which often fails to take into account perpetration frequency, doesn’t consider harassment of victims as a related offence and does not compare intentional and non-intentional encounters with this content. Secondly, we are concerned with intervention design, both in the long-term efficacy of interventions and the need to draw from stakeholder experience in design. Finally, we view the social nature of deepfake abuse as an understudied aspect of the harm which required both empirical study and targeted solutions.

Traces of Abuse: How Generative AI Impacts Image-Based Sexual Abuse (IBSA) Investigations

Jasmin Wyss (Presenter, Ruhr-Universität Bochum), Ivy Turk (Ruhr-Universität Bochum), Anna Neumann (University Duisburg-Essen), and Rebekah Overdorf (Ruhr University Bochum)

The introduction of generative AI (GAI) into the workflow of image-based sexual abuse (IBSA) only worsened the ease of creation and distribution, victimizing more people than ever. We outline how the introduction of generative AI (GAI-IBSA) impacts the creation of traces and the type of reasoning they allow. We illustrate the impact by comparing the forensic traces available in four different IBSA scenarios. We discuss the impacts on the (possibility of) investigation, arguing that the advent of generative AI overall benefits abusers by making perpetration easier, and the perpetrator harder to trace.

Enclosures of the Body: Platform Capitalism, Primitive Accumulation and Image-based Sexual Abuse in Singapore

Raag Sudha Sanjay (National University of Singapore)

Image-based sexual abuse (IBSA) – the non-consensual capture, creation, and distribution of sexually explicit imagery – persists on major corporate advertising platforms (such as Meta and TikTok) at significant scale despite legal prohibition, explicit platform policy, and documented reputational consequences for non-compliance. I argue that the dominant regulatory account, which treats IBSA as a governance deficit awaiting legal correction, misdiagnoses the phenomenon. Drawing on Marxist-feminist political economy, this research proposes that systemic under-moderation of IBSA is a structural feature of the valorisation logic governing advertising-based platforms, not an aberrant failure of governance. I develop the concept of the spectral body to theorise the structural position of the IBSA victim within the platform's valorisation circuit: expropriated from her intimate imagery, present as a data-generating object, absent as a consenting subject. Content moderation is theorised as faux frais – a cost of circulation minimised structurally because it generates no surplus value. Singapore serves as the primary empirical site; being both the regional headquarters of Meta and TikTok and the producer of Southeast Asia's most sophisticated IBSA regulatory architecture, it constitutes the analytically strongest case for testing whether under-moderation persists as a structural tendency even under conditions of maximum regulatory investment and high state capacity. The research aims to apply critical discourse analysis to platform governance documents and Singapore's legislative corpus, interrogating how each constructs IBSA and whether either framework engages with the accumulation logic that sustains it. The findings carry implications for feminist digital rights scholarship, the limits of legalism in Asian developmental states, and the governance of global platform capitalism across the Asia-Pacific.

---