Comparing Human and AI-Driven Nominal Group Techniques for Validating Tourism Sustainability Indicators

Reaching a consensus among experts on sustainability indicators is essential for evidence-based tourism management. However, traditional methods often face logistical and resource constraints. This study explores an alternative by comparing the conventional human-led Nominal Group Technique (NGT) with an AI-driven multi-agent version to validate tourism sustainability indicators for Barra and Costa Nova, Portugal. The goal is to assess whether AI simulations can replicate the depth and rigour of human expert discussions while enhancing efficiency and scalability.

The study examined how human experts and AI-simulated agents reached consensus on economic, environmental, and sociocultural indicators. Six AI agents were modelled on expertise in tourism, environment, economics, and culture to mirror the human expert panel. Both groups followed the same two-round NGT process to assess 73 indicators. In the second round, participants revisited indicators lacking consensus and considered new suggestions. Both humans and AI agents also proposed improvements to the monitoring framework and specific indicators. Comparative counts, based on decision outcomes by domain and round, and narrative synthesis analyses of differences in suggestions and justifications were used to examine agreement levels, reasoning patterns, and final indicator selections.

Both groups agreed on about a quarter of the indicators, with the most substantial alignment in environmental monitoring. Differences were more pronounced for economic and sociocultural indicators, where contextual interpretation played a key role. In the second round, human experts agreed on context-sensitive sociocultural indicators, while AI agents showed greater consistency in technical and environmental measures.

Findings suggest that the AI-driven multi-agent NGT offers a scalable and efficient tool for preliminary screening of sustainability indicators. Beyond methodological innovation, hybrid approaches combining AI efficiency with human contextual insight have practical potential for continuous monitoring, rapid reassessment, and integration into adaptive tourism management and policy design.