Thesis title: Transforming Tourism Sustainability Through Artificial Intelligence: A Multidimensional Framework for Overtourism Prediction and Mitigation
This research develops a multidimensional framework leveraging artificial intelligence to predict and mitigate overtourism, addressing the limitations of traditional tourism management approaches. Through a systematic review of 150+ academic publications (2010-2025), the study demonstrates how AI models achieve 89-94% accuracy in tourism flow prediction, significantly outperforming conventional methods (65%). The research introduces the Overtourism Risk Index (ORI), an innovative metric integrating demographic pressure, spatial congestion, ecological vulnerability, and infrastructure strain. Case studies of Rome (ORI: 0.48, moderate risk) and Venice (ORI: 0.97, critical risk) reveal how visitor volume alone inadequately determines sustainability: Rome manages 51.4 million annual visitors through spatial dispersion capacity, while Venice with 25-30 million exhibits extreme vulnerability due to geographic constraints. AI system implementation demonstrated 23% reductions in major attraction congestion and 18% decreases in transport-related emissions. The proposed framework transforms tourism management from reactive to proactive, enabling evidence-based interventions. The research critically addresses privacy implications, proposing differential privacy techniques and learning approaches. This work establishes theoretical and practical foundations for sustainable destination governance in the artificial intelligence era.