In a 2019 article “GeoAI: spatially explicit artificial intelligence techniques for geographic knowledge discovery and beyond”, we speculated about a potential GeoAI moonshot to develop an autonomous GIS analyst that can perform simple spatial data science tasks entirely by itself. We suggested that our community may benefit from a rallying around a common goal, also to better measure progress by developing a reproducible benchmark. We assumed that GI bots that can master common tasks performed by entry-level/junior GI analysts could be developed by 2030.
Given recent progress in AI and machine learning, and with the rise of foundation models we now believe that systems with an interaction style similar to ChatGPT may arrive sooner, maybe even in 2025. Such systems will be able to react to a user's written or spoken instructions and then independently search data, process data, perform multi-step GIS operations, and then provide appropriate visualizations back to the user together with an answer to the user's request. While still in their infancy, we are already seeing first prototypes in development using a combination of LLM and RAGs.
While these systems will truly bring spatial data science to the masses, developing autonomous GI bots is not without risks, and the increasing tendency to connect ChatGPT-like systems to IoT devices is worrying. We believe the best next step for our spatial data science and GeoAI communities is to develop a benchmark and competition series that supports experiments and fruitful competition around developing autonomous GIS bots while at the same time being guided by strong ethical guidelines and transparency. Hence, we want to understand better which tasks can already be performed right now with existing technology, which tasks require geo-specific bots and geo-foundation models, which workflows will require substantial human input for the years to come, and whether there are substantial roadblocks in reasoning and representation.
The workshop will provide an interactive forum to engage in discussions, shape the research directions, and disseminate state-of-the-art research on Autonomous GIS Bots. Example topics include, but are not limited to, the following:
The workshop seeks high-quality full (8-10 pages) and short (4 pages) research papers as well as vision statement (2 pages) that have not been published in other academic outlets and are not concurrently under peer review. Interested participants should submit a paper in the ACM format. Once accepted, at least one author is required to register for the workshop and the ACM SIGSPATIAL conference, as well as attend the workshop to present the accepted work which will then appear in the ACM Digital Library.
To be announced.
To be announced.