Supporting Boundary Management in Young People’s AI Companion Interactions

Overview

We invite applications for a PhD position within an interdisciplinary research team examining how young people engage with AI companion chatbots (e.g., Character.AI, Replika) and how they manage risks and boundaries in these interactions.

AI companions have grown in popularity since the COVID-19 pandemic and are increasingly used by young people seeking connection. This trend raises important concerns, including exposure to harassment, misinformation, and self-harm, as well as the potential impact on human relationships when reliance on AI companions becomes significant.

This project aims to generate new knowledge and design interventions that help young people establish safe and healthy boundaries with AI companions. While boundaries are well understood in human relationships, little is known about their role in human–AI interactions, where risks are borne entirely by the human user.

This research topic is connected with ongoing research on generative AI in the Digital Wellbeing Lab.

Research activities

As a PhD student in this project, you will engage in the following research activities:

• Qualitative: Use diaries and semi-structured interviews to explore how young adults manage boundaries and risks with AI companions.
• Co-Design: Collaborate with young adults to develop interventions that support boundary management (e.g., tools to flag boundary violations, set conversation limits, and stop harmful interactions). This includes prototyping and testing solutions for online environments.
• Quantitative: Conduct experimental studies to evaluate intervention effectiveness, focusing on outcomes such as engagement, perceived safety, and risk reduction.

Outcomes

The project intends to generate the following outcomes:

• The first Australian study on how young adults can be supported to manage risks in interactions with AI companions.
• New theoretical insights into boundary management in one-sided relationships with AI agents.
• Methodological advances through co-design approaches that enable non-AI experts to identify safety risks and develop effective interventions.

Skills and experience

We are looking for a student with a recent Masters or Honours degree in Human-Computer Interaction, Psychology, or a related field. This project requires strong skills in qualitative and/or quantitative research methods, and an interest in participatory research with young adults (16-25) to explore emerging AI technologies.

Scholarships

You may be eligible to apply for a research scholarship.

Explore our research scholarships

Keywords

• AI companions
• Replika
• Character.ai
• AI safety
• generative AI
• human-AI interaction
• human-AI relationship
• Human-computer interaction (HCI)
• human-computer interaction
• boundary management
• relational harm
• AI harm
• online safety
• Large language models (LLMs)
• parasocial relationships
• co-design
• participatory design
• chatbot safety

Contact

Contact the supervisor for more information