Development of AI geo-agents for retrieval, analysis, and visualization of spatial data

In recent years, the increasing availability of large language models (LLMs) -such as Claude 3 or GPT 4 -combined with the need to process ever larger volumes of spatial data, has opened new perspectives for Geographic Information Systems (GIS) and cartography.At the same time, many research communities wonder whether, in the coming years, we will achieve the state of so-called Artificial General Intelligence (AGI), also referred to as the singularity point, beyond which it is difficult to predict further directions of development.In the context of geoinformatics, the question is whether we can construct a fully autonomous GIS solution in which an intelligent agent would be capable of passing a "GIS Turing test" (Janowicz, 2020).In other words, a user formulates a query (e.g., for selecting the best location for solar installations), and the agent independently acquires the appropriate data layers from spatial data infrastructures (SDI), performs advanced analyses (e.g., solar irradiance analysis), and returns a map indistinguishable from work done by a human expert.The concept of a "GIS artificial analyst," proposed by Janowicz (2020), aims to develop by 2030 software capable of independently solving complex geoinformation problems.