Current Research
Below are my current and active research projects. For completed and published manuscripts, see Publications.
Artificial Social Influence
It is becoming evident that interactions with artificial intelligence (AI) based agents can positively influence human behavior and judgment. However, studies to date focus on text-based conversational agents (CA) with limited embodiment, restricting our understanding of how social influence principles apply to AI agents. We address this gap by leveraging the latest advances in AI (language models) and combining them with immersive virtual reality (VR) to build VR-embodied conversational agents (VR-ECAs) that can naturally converse with humans.
The first study implementing this research paradigm specifically examined the effect of human-agent similarity (via gender matching) on social, behavioral, and biobehavioral outcomes:
Lim, S., Schmälzle, R, & Bente, G. (2024). Artificial social influence via human-embodied AI agent interaction in immersive virtual reality (VR): Effects of similarity-matching during health conversations. arXiv. https://doi.org/10.48550/arXiv.2406.05486
This work was nominated for the Google PhD Fellowship for the 2024-2025 school year (results pending).
Examination of Message Effects via Immersive VR
We are constantly exposed to persuasive messaging in our day-to-day lives, yet it is difficult for researchers to precisely measure how much attention people pay to these messages:
- With stationary eye-trackers, it is difficult to study how messages attract attention in a natural-real-world setting.
- In controlled laboratory settings, it is difficult to achieve ecological validity and account for real-life situational factors.
Immersive virtual reality (VR), integrated eye-tracking capabilities, is a promising platform that allows for the examination of message effects in simulated, but realistic environments (Bonneterre, 2023; Schmälzle et al., 2023). In addition, the advancements in generative AI tools allow us to augment our message development processes.
Building on Schmälzle et al.’s (2023) billboard paradigm, the CARISMA Lab’s most recent study had participants drive through a virtual city with 40 billboards distributed throughout. We then examined how the participants’ task condition (sale sign counting vs. free driving), billboard location (start vs. middle vs. end), and visual message design (less vs. more attention grabbing) influenced their viewing behavior (i.e., fixation count and gaze duration) and memory (recall and recognition).
This study received an honorable mention (best poster) at Kentucky Conference on Health Communication (KCHC) 2024.
References
Bonneterre, S. (2023). Evaluating smoking prevention campaigns in virtual reality: An experimental ecological approach (Doctoral dissertation, Université de Nanterre-Paris X).
Schmaelzle, R., Lim, S., Cho, H. J., Wu, J., & Bente, G. (2023). The VR billboard paradigm: Using VR and eye-tracking to examine the exposure-reception-retention link in realistic communication environments. PlosOne, 18(11), e0291924.