In “Emerging trends in geospatial artificial intelligence (geoAI)”, VoPham et. al explain the emergence of geoAI as a new research field combining concepts, methods and innovations from various fields, such as spatial science, artificial intelligence (AI), data mining and high performance computing, and give examples of recent applications in real-life situations. The fusion between AI and GIS helps us obtain more accurate representations compared to traditional methods given the ability to make use of spatial big data.
As mentioned in the article, geoAI has the ability to revolutionize remote sensing, with the potential to more accurately recognize earth features. Slight differences in the spectral response of a pixel could be detected by an algorithm trained to detect these ever so small differences, which could help detect and respond to forest fires more rapidly for example. A research project I worked on last year aimed at assessing the extent of the Fort McMurray forest fire of 2016, and although the results were extremely similar to what had been obtained by official government sources, the use of geoAI could have overcome the limitations of the NDVI and NBRI indices used.
As with any new emerging scientific field, it will be interesting to see how and to what geoAI will be applied to next. An example would be spatial Agent-based modelling (ABM), which aims to simulate the actions of specifically defined agents in space, which could highly benefit from geoAI and the input from spatial big data. Geographical ontologies could also be redefined by deep learning, which could conceptualize things differently from the way we currently do.