How generative AI shapes customer input on video platforms

Generative AI has profoundly influenced numerous industries. Nevertheless, its impact on video platforms, particularly regarding the interaction between textual and visual modalities, remains underexplored. Drawing on cue consistency theory and approach-avoidance motivation theory, the present research employed three experiments and sentiment analysis to investigate how generative AI output modalities shape customer input on video platforms. The findings reveal that generative AI textual modality alone can influence customer input through novelty and cynicism. However, this effect is substantially shaped by cross-modal congruence in multimodal environments. Normal text demonstrates the greatest alignment with logical images, whereas poetic text achieves the highest congruence with emotional images. This research contributes to the theoretical understanding of generative AI's influence on video platforms and provides practical recommendations for designing generative AI to improve user retention.