Ethics in data-centric AI for geosciences: a double-edged sword

The Geosciences, especially Earth Observation or Remote Sensing research, turn increasingly to Artificial Intelligence to manage the petabytes of data collected by an ever-increasing number of satellites. EO/RS are indeed heavily dependent on data, and scientists are on the hunt for more data, better data, and more representative and diverse data to make their AI models accurate and reliable across diverse regions of the globe. However, can this data-centric AI for EO be a double-edged sword also leading to ethical concerns? This paper highlights major ethical guidelines in the context of AI and big data and provides concrete examples of how these might play out in the context of both small and very large AI4EO research projects such as the Digital Twin Earth (DTE).