From 7 to 8 May 2024, a workshop was held in Frankfurt am Main to examine the application of artificial intelligence (AI) in geoscience. It was organised by the Federal Agency for Cartography and Geodesy (BKG) in cooperation with the Federal Environment Agency (UBA).

The workshop provided a platform for professional exchange on successful AI applications in the analysis of geodata. Topics such as remote sensing-based monitoring, land cover changes and efficient object recognition were discussed in various sessions. Questions such as “Can the health status of a tree be determined with laser point clouds?” or “How can flood protection be improved using remote sensing data and AI?” were addressed. Speakers presented innovative approaches and projects that use AI methods to analyse environmental data, monitor land cover changes and efficiently extract objects from geodata. In a poster session, PhD student Rewanth Ravindran presented his work on “Segment Anything Modle (SAM) for Remote Sensing”.

A particular highlight was the panel discussion in which Prof Dorota Iwaszczuk and other experts discussed the potential and limitations of geo-AI. Challenges and opportunities for the integration of AI in geoscience were highlighted. The topics of AI's hunger for data, the need for good training data, data security and the generalisation of methods dominated the discussion. For the future, the focus was clearly placed on stronger collaboration.

“The best projects are not done alone.”

The workshop offered participants the opportunity to gain new insights, take away important ideas for future projects and network with each other. The discussions showed that the integration of AI into geoscience offers promising developments and innovative solutions for current challenges.