Geographic Information is now undergoing a profound shift from traditional physics-based modelling to advanced data-driven AI methods for better processing of complex geospatial data. 

While AI can transform traditional maps and remote sensing imagery into spatial insights for decision-making and unmanned systems, significant challenges exist, such as geospatial training data diversity, reproducibility of Geospatial AI (GeoAI) models, and inconsistencies in model prediction results. To overcome these hurdles, the community urgently needs a comprehensive standardization framework covering GeoAI data/model/services. 

Recognizing this critical need, ISO/TC211 has emerged as a pioneering force in addressing these standardization challenges through systematic and comprehensive efforts. The standard ISO 19178-1 Geographic information — Training data markup language for artificial intelligence specifically targets the standardization of GeoAI training data, significantly facilitating the sharing and interoperability of AI training samples. 

Building upon this foundation, the ISO/PWI 19182 project represents an undertaking to systematically incorporate GeoAI technologies within broader geographic information standards, which can help establish the groundwork for a unified standardization framework. 

Existing geospatial standards such as ISO 19115 metadata and 19157 data quality parts will be leveraged for consistent descriptions of metadata, provenance, quality of GeoAI data and models, which will improve the understanding and responsible use of GeoAI.

Through the standardization initiatives, ISO/TC211 is strategically positioning GeoAI to evolve from traditional spatial data infrastructures (SDI) into AI-ready SDI enriched with GeoAI data, models, and services. 

By establishing foundational standards around GeoAI infrastructures in the future, ISO/TC211 is ensuring that GeoAI can transition into a reliable, trustworthy technology that truly serves the global community.