Deepfakes are AI-generated images, videos, and voices depicting real people. Initially a fringe source of entertainment, deepfakes have evolved into a mass-market technology with consequences for evidence, identity, and mental health, and cases of deepfake-based abuse include sexualization, financial fraud, defamation, spread of disinformation, and identity theft. My work approaches deepfakes from three angles: How well humans can detect them and how to improve detection, what psychological and social harms they cause, and how this technology can be used constructively in clinical contexts. It also opens up new questions on the experiential dimensions made possible by deepfakes: What happens to us when we see a deepfake of ourselves, in a situation we never experienced before?

Current research:
- DeepSelf (IFORES Career Kickstart, 2026): An experimental study on psychophysiological responses of deepfakes of oneself, including sexualized deepfakes, with implications for victimization and clinical care.
- Empirical and review work on how humans detect deepfakes and how this performance can be improved.
- Research on the perception of and response to AI-generated social media figures like AI-generated influencers or physicians.
- Conceptual research on the clinical applications of deepfake technology.

Selected contributions:
-  Diel, A., Lalgi, T., Schröter, I. C., MacDorman, K. F., Teufel, M., & Bäuerle, A. (2024). Human performance in detecting deepfakes: A systematic review and meta-analysis of 56 papers. Computers in Human Behavior Reports, 16, 100538.
-  Diel, A., Lalgi, T., Mellis, F. S., Teufel, A., & Bäuerle, A. (2025). The harm of deepfakes: a scoping review of deepfakes’ negative effects on human mind and behavior. AI & SOCIETY, 1-17. 

General-purpose chatbots are now used by millions of people for emotional support, often outside any clinical framework and without sufficient safety guardrails designed for mental health vulnerability. The field and media often oscillate between optimistic accounts of flexible, constantly available mental health support and alarming case report of 'AI psychosis', emotional dependence, or chatbot-induced suicides. Within the framework of a Psychopathology-AI-interaction, I investigate how mental health and use of LLM-based chatbots interact with one another, focusing on vulnerability, interaction patterns, and chatbot design features.  

Current research:
- A Scoping Review on the harms of LLM-based chatbot use for mental health
- Routine data analysis on how patients use LLM-based chatbots for mental health
- Vignette and patient-chatbot interaction research
- Development of user guidelines for the use of LLM-based chatbots for mental health

Digital mental health interventions such as apps, online programmes, AI-based tools, and robotic systems, find increasing use in outpatient mental health treatment. However, their integration into inpatient treatments is relatively underdeveloped. My work investigates how e-mental health (EMH) tools can be implemented into inpatient care across the inpatient journey, and how hurdles of implementation (such as adherence) can be challenged.

Current research:
- Evaluation of the efficacy and acceptance of EMH-based inpatient treatment.
- Development and evaluation of AI-based EMH tools, such as AI-generated mindfulness exercises (MindfulAI).
- Implementation of social robots in mental health inpatient care.
- How to overcome hurdles of EMH tool application (e.g., adherence to EMH programmes, acceptance of robotic or AI-based agents).

Selected contributions:
- Diel, A., Schröter, I. C., Frewer, A. L., Jansen, C., Robitzsch, A., Gradl-Dietsch, G., ... & Bäuerle, A. (2024). A systematic review and meta analysis on digital mental health interventions in inpatient settings. NPJ Digital Medicine, 7(1), 253.
-  Diel, A., Schröter, I. C., Robitzsch, A., Jansen, C., Teufel, M., & Bäuerle, A. (2025). E-Mental Health Interventions in Inpatient Care: Scoping Review. Journal of Medical Internet Research, 27, e65947.