Detecting Creative States From Emotional Cues: Insights From Speech With Focus on Text Modality

Creativity and emotion are fundamental aspects of human expression, interaction, and innovation, shaping various facets of modern society. This paper explores the intersection of technology, human creativity and emotional expression by proposing a novel approach to detecting creative states through the analysis of emotional cues derived from speech. To validate the findings, a linguistic prompt creativity test was conducted, incorporating both human evaluations and AI-driven assistance as the co-collaborative approach, using the Torrance Test of Creative Thinking (TTCT) as the evaluation criteria. The study analysed 50 creative speech samples from 25 participants, each exposed to happy and sad video stimuli to evoke distinct emotional responses. The results reveal a correlation between emotional states and creativity. A paired t-test revealed a difference in creativity scores between the happy and sad conditions (t(24) = 2.46, p = 0.021), with higher creativity observed in the happy condition. In 80% of the participants, higher creativity levels were observed when they experienced positive emotions and happiness, highlighting the significant influence of emotional context on creative expression and creative states and also aligning with the psychological research on exploring the effects of emotions on creativity.