Abstract
We propose Collab-REC, a multi-agent framework designed to counteractpopularity bias and enhance diversity in tourism recommendations. In oursetting, three LLM-based agents -- Personalization, Popularity, andSustainability generate city suggestions from complementary perspectives. Anon-LLM moderator then merges and refines these proposals via multi-roundnegotiation, ensuring each agent's viewpoint is incorporated while penalizingspurious or repeated responses. Experiments on European city queries show thatCollab-REC improves diversity and overall relevance compared to a single-agentbaseline, surfacing lesser-visited locales that often remain overlooked. Thisbalanced, context-aware approach addresses over-tourism and better aligns withconstraints provided by the user, highlighting the promise of multi-stakeholdercollaboration in LLM-driven recommender systems.