AI Companionship: Understanding Harms in Human–AI Relationships
Contact: Tarini Saka
AI companions are increasingly becoming part of people’s social and emotional lives, offering opportunities for connection, intimacy, and support. However, these emerging relationships also raise important questions about potential harms to users. This research explores the risks associated with AI companionship, focusing on how harmful experiences emerge, how users perceive and respond to them, and what forms of support or recourse are available. Potential topics include emotional dependency, manipulation, harmful interaction patterns, and platform accountability. The project aims to develop a better understanding of harms in human–AI relationships and contribute to the design of safer, more responsible AI companion systems.
Beyond the topics offered on this page, you can always propose your own topic. Please ground your proposal in my group’s recent publications and current research directions. Email me using the application template by RUB HCS Chair. You may also include your academic transcript, resume, and other relevant materials.
Before I formally accept an application, I will usually ask you to write an exposé based on our initial discussions. The exposé should address the following points in roughly three pages:
I will also provide feedback on the exposé. Once it is final, we can sign it as our thesis agreement and you can start the registration process at the examination office.
Email |
Google Scholar |
LinkedIn |
Site built with Jekyll, template here