Paula Friedrich's Master Thesis is accepted in the Springer Best Masters Series

In her master’s thesis, Paula Friedrich investigates how embodied artificial intelligence can be used ethically and effectively in virtual reality–based mental health care. Against the backdrop of a global shortage of mental health professionals and rising numbers of affected individuals, her work addresses the potential of AI and VR to support and extend existing therapeutic services.

In her thesis, she first examines mental health patients’ key concerns regarding the use of AI and VR in therapy through an online questionnaire. The results highlight data autonomy and transparency about the AI system as the most important factors, closely aligning with established ethical principles in human-centered AI.

Building on these findings, Paula Friedrich conducts an in-person VR study to analyze how transparency and data autonomy influence self-disclosure and trust toward an embodied AI therapist. Her results show that higher levels of data autonomy lead to increased self-disclosure in a VR-based questionnaire, although this effect diminishes when participants are asked more sensitive questions—reflecting patterns known from human therapist–patient interactions. Transparency does not affect self-disclosure, and neither transparency nor autonomy significantly influence trust in the AI. However, her study reveals a gender effect on trust, with women reporting higher trust levels than men.

Overall, in her thesis, Paula Friedrich contributes valuable empirical insights to the ethical debate on AI in mental health care. Her work adopts a human-centered AI perspective, sheds light on patients’ concerns, and takes an important step toward translating abstract ethical principles into concrete technical design implications for embodied AI systems in therapeutic contexts.

The thesis can be accessed here