Abstract for: Exploring how features of epistemic trust influence system thinking as stakeholders engage with system dynamics and AI

Epistemic trust (ET) may affect stakeholder learning and decision-making in social learning contexts of group model building (GMB). Stakeholder interactions with system dynamics models using artificial intelligence may be affected by ET, i.e., trust given to information sources based on social cues like expertise and interpersonal attunement. We explore how socioemotional features of interactions, like selective trust in the communicated knowledge about systems during the conduct of these methods, influence stakeholder learning. We apply qualitative mapping to a literature review on epistemic trust methods (in therapy, in empirical measurement). We do so to generate novel system insights about how processes in epistemic trust may influence stakeholder engagement and learning with commonly used practices such as GMB as well as novel system dynamics modeling practices involving generative artificial intelligence (AI). We report on what is learned when applying literature on epistemic trust when considering specific examples of stakeholder engagement in system dynamics methods. We present causal loop diagrams (CLD) to understand how socioemotional features like capacity to reliably detect communication cues may reinforce system-based learning when deploying system dynamics methods with community stakeholders and inclusive of AI. We present a preliminary CLD based on an initial literature search that combined experimental, observational, self-report, and neuroimaging studies about relevant socio-cognitive processes involving epistemic trust. These causal hypotheses demonstrate how ET may influence information processing and participant engagement with system thinking methods in real-time, as well as in a context of generative AI model use. Implications for how current system dynamics methods could incorporate ET when engaging with stakeholders are considered.