AI-Mediated Emotional Expressivity in Music Animation Pedagogy: A Cross-Domain Study of Affective-Cognitive Integration

Resumo

The aim of this study is to test its utility in complementing music animation pedagogy through cross-domain research: the intersection of auditory tonalities and visual synchronicities, highlighting the limitations of traditional approaches to emotional literacy. Affectiva and OpenFace, chosen for their ability to quantify and articulate nuanced affective states, were used among 149 students (their demographics, differences in prior learning, and cultural context form the basis for interpretive comparisons), revealing a marked shift in 'emotional congruence' (+35% alignment of sound and image elements), self-awareness (+30%) and articulation accuracy (+40%). The dichotomy of pre-intervention impairments (65% failure to elicit consistent emotional intention) and post-intervention transformations (88% success in conveying expression)—reinforced by the emergence of tonal structures—suggests an epistemic redefinition of emotional intelligence itself, mediated by AI. The methodological design—a mixed method enriched by qualitative analysis and quantitative analysis—emphasises the potential of such tools, but reveals inherent limitations: biases embedded in the algorithms (contextual nuances that are often circumvented by the universal logic of AI), regional peculiarities of pedagogical perception and different levels of participants' basic 'emotional literacy'. The resonance of this study—with its implications for cognitive-affective integration—lies in its findings in disrupting established binaries in the emotion/cognition, teaching/learning, and human/algorithmic processes relationships. By considering AI not as a replacement but as a 'mediator' of affective development (a conceptual bridge linking latent potential to explicit expression), this research highlights the untapped synergy of computational methodologies and artistic praxis.