Answers are grounded in this paper and cite their sources.
The present RCT investigated the short-term effectiveness of a voice-based conversational Artificial Intelligence (AI) agent (Maya and Miles) compared to a traditional psychoeducational video in reducing state anxiety among undergraduate students during exam periods. Sixty participants (aged 18–22) who reported elevated levels of state anxiety on the State-Trait Anxiety Inventory (STAI-S) were randomly assigned to one of the two conditions: the experimental group (n = 30) engaged in a 12–15-min interaction with a conversational AI agent (Maya or Miles), while the control group (n = 30) viewed a 15-min TED Talk on anxiety management. Pre- and post-intervention anxiety scores were recorded. A two-way mixed ANOVA revealed a significant main effect of time and group (F = 73.17, & 28.28 p < .001), respectively, indicating a reduction in anxiety across both groups. The group × time interaction was not significant, indicating no statistical evidence that one intervention was more effective than the other. However, effect size analysis revealed large within-group reductions in anxiety for both the conversational AI group (d = −1.12) and the video group (d = −1.10), underscoring the strong impact of both interventions. These findings suggest that both interactive and non-interactive digital interventions may serve as viable short-term tools for managing state anxiety in academic contexts. Future studies should explore longer-term effects and the role of user expectations and engagement styles in shaping outcomes.
No takes yet. Share an insight, caveat, or question.
Pareek et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: