Context-Aware Background Music Recommendation System for Everyday Conversations Using Generative AI

Abstract

With the increasing use of music software, many users prefer integrating

background music into everyday conversations to enhance emotional interaction.

However, traditional music recommendation systems fail to capture the subtle

contextual nuances of conversations in real-time for BGM recommendations. This

paper proposes a novel system that transcribes spoken dialogues into text and

leverages the contextual understanding capabilities of AI large models to match

conversation content with real-life interaction scenarios, recommending appropriate

BGMs from a curated dataset. In this way, the system provides dynamic BGM

recommendations for conversations, which not only enhance user immersion but also

open up new avenues for further exploring the role of AI in enhancing human

experiences through contextual music recommendations.