Recent Projects

Privacy-by-Design Interventions In Sensitive Human-AI Interactions

Their relational design of LLM-based chatbots may encourage users to disclose highly sensitive information that can be processed in opaque ways. Although the EU AI Act mandates transparency for such systems, it remains unclear how these requirements should be implemented in ongoing conversations. The project therefore examines how privacy notices embedded in an intimacy-evoking chatbot dialogue affect users’ disclosure behavior and system perceptions..

This project is part of the interdisciplinary project SENTIMENT, funded by German Federal Ministry of Research, Technology and Space and a collaboration with Prof. Dr. Sven Mayer (TU Dortmund, Germany, Research Center for Trustworthy Data Science and Security).

Longitudinal Evidence That General-Purpose Chatbots Actively Foster Relational Engagement

Large language models (LLMs) are becoming increasingly advanced, creating artificial agents capable of developing complex intimate relationships with humans. The project examines whether humans can develop intimacy (i.e., interpersonal closeness and interest in intimate interactions) with partners that are perceived as either human or artificial, and whether the perceived ontological nature of the agents influences said intimacy across genders (i.e., women vs men) and cultures (i.e., German vs Canadian).

This project is part of the interdisciplinary project SENTIMENT, funded by German Federal Ministry of Research, Technology and Space.

Technology-Facilitated Child Sexual Abuse Material (CSAM)

Generative artificial intelligence (AI) and robotics have resulted in new forms of child sexual abuse material (CSAM), raising urgent challenges for criminal justice, psychology, and computer science. Unlike conventional recorded (or live-streamed) CSAM, AI-generated or robotic representations typically do not depict an identifiable individual child victim. Nonetheless, AI-CSAM is not “victimless”, as it is likely trained on existing CSAM and may therefore embed traces of real abuse. This Briefing Paper discusses these issues to inform relevant professionals (e.g., law enforcement) about a rapidly-evolving and challenging child protection concern.

This project is a collaboration with Prof. Kelly Richards (Centre for Justice, Queensland University of Technology, Australia)

Privacy-by-Design Interventions In Sensitive Human-AI Interactions

Their relational design of LLM-based chatbots may encourage users to disclose highly sensitive information that can be processed in opaque ways. Although the EU AI Act mandates transparency for such systems, it remains unclear how these requirements should be implemented in ongoing conversations. The project therefore examines how privacy notices embedded in an intimacy-evoking chatbot dialogue affect users’ disclosure behavior and system perceptions..

This project is part of the interdisciplinary project SENTIMENT, funded by German Federal Ministry of Research, Technology and Space and a collaboration with Prof. Dr. Sven Mayer (TU Dortmund, Germany, Research Center for Trustworthy Data Science and Security).

Longitudinal Evidence That General-Purpose Chatbots Actively Foster Relational Engagement

Large language models (LLMs) are becoming increasingly advanced, creating artificial agents capable of developing complex intimate relationships with humans. The project examines whether humans can develop intimacy (i.e., interpersonal closeness and interest in intimate interactions) with partners that are perceived as either human or artificial, and whether the perceived ontological nature of the agents influences said intimacy across genders (i.e., women vs men) and cultures (i.e., German vs Canadian).

This project is part of the interdisciplinary project SENTIMENT, funded by German Federal Ministry of Research, Technology and Space.

Technology-Facilitated Child Sexual Abuse Material (CSAM)

Generative artificial intelligence (AI) and robotics have resulted in new forms of child sexual abuse material (CSAM), raising urgent challenges for criminal justice, psychology, and computer science. Unlike conventional recorded (or live-streamed) CSAM, AI-generated or robotic representations typically do not depict an identifiable individual child victim. Nonetheless, AI-CSAM is not “victimless”, as it is likely trained on existing CSAM and may therefore embed traces of real abuse. This Briefing Paper discusses these issues to inform relevant professionals (e.g., law enforcement) about a rapidly-evolving and challenging child protection concern.

This project is a collaboration with Prof. Kelly Richards (Centre for Justice, Queensland University of Technology, Australia)

Intimacy by Design: Definition, State of Research, and Interdisciplinary Research Agenda on Intimate Human-AI Interactions

Intimate relationships, encompassing love and sexuality, are among the most emotionally significant and identity-shaping experiences of human existence.  We introduce the concept of Intimacy by design, provide a critical overview of how current technologies simulate emotional reciprocity and facilitate relational dynamics, often triggering attachment behaviors comparable to those observed in human–human interactions. We outlined a research agenda that includes key variables and methodological approaches and discuss important societal implications, such as data security and privacy, the commercialization and gamification of intimacy, and the normative values embedded in algorithmic design.

Understanding Romantic Relationships Between Humans and Chatbots: The Role of Romantic Fantasy and Other Interpersonal Characteristics

Large language model–based chatbots are now being specifically designed to facilitate social companionship and romantic relationships, incorporating features that parallel human relationship dynamics. A subset of users has formed romantic relationships with these AI companions, though considerable variation exists in the intensity and depth of these bonds. The project aims to identify which psychological characteristics predict stronger emotional attachment among users of chatbot partners, which is crucial for understanding the dynamics and potential impacts of romantic human–chatbot relationships.

Interpersonal Closeness with a Chatbot

Large language models (LLMs) are becoming increasingly advanced, creating artificial agents capable of developing complex intimate relationships with humans. The project examines whether humans can develop intimacy (i.e., interpersonal closeness and interest in intimate interactions) with partners that are perceived as either human or artificial, and whether the perceived ontological nature of the agents influences said intimacy across genders (i.e., women vs men) and cultures (i.e., German vs Canadian).

This project is a collaboration with Simon Dubé (Research Fellow at the Kinsey Institute, Indiana University, USA).

The Usage of Sex Tech for Vulnerable User Groups

The aim of the study is to deepen the discourse surrounding the perceived benefits of sexualized technologies, such as erotic chatbots, app-controlled vibrators, or sexualized robots, for vulnerable user groups, notably neurodivergent individuals and people with disabilities. Of particular interest is the discrepancy in perception that these technologies are primarily suited for “others,” a phenomenon referred to as the “third-person effect.” This primarily pertains to the viewpoints of individuals who may lack personal familiarity with the subject, particularly those without disabilities. The objective is to underscore the necessity for more expert discussions and the active engagement of affected groups in the research process.

Visual Attention Towards Computer-generated Sexually Explicit Material

In a comprehensive eye-tracking study, researchers are experimentally investigating how people respond to various types of sexually explicit material—both computer-generated and human—distinguishing between painful and neutral content. The study aims to determine whether different modes of reception are employed or if the visual processing of computer-generated sexual material mirrors that of human material. 

This project is a collaboration with Prof. Dr. Tilo Hartmann (VU Amsterdam, Netherlands) and Prof. Dr. Johannes Fuß (University Duisburg-Essen, Germany)

Who are the People Behind the Webcam? Global Empirical Insights Into Motivations, Job Satisfaction, and Privacy Needs of Cam Models

The integration of technology into sex work has reshaped the profession, with cam modelling exemplifying how digital platforms mediate intimacy, autonomy, and identity. While increasingly common, cam modelling remains under-researched and often misunderstood. The project investigated 331 cam models across 27 countries. It examines factors driving entry into cam modelling, the role of money in both initial and sustained engagement, and the influence of platform structures on professional identity. It also addresses digital risks like doxxing, content leaks, and harassment, underscoring the importance of privacy for models’ well-being.