AI onboard satellites: here’s how it can help

AI will assist satellites in optimizing the response to natural disasters, says the European Space Agency (ESA).

When floods, fires, or landslides occur, satellite data plays an important role in disaster management. In places beyond the reach of conventional networks, satellites help to provide information and maintain communication when land-based infrastructure might be down.

However, data processing has been challenging for scientists. Satellites collect huge volumes of data from various inputs, which makes it hard to process and send to the first responders in a timely manner. During a natural disaster, accurate information is needed within minutes.

Here is how AI can help. During Milan’s 75th International Astronautical Congress, the ESA announced a newly launched project called Ciseres, which aims to optimize data processing to assist first responders.

AI will be onboard the satellite, at the front lines, scanning the flood of satellite data to pinpoint crucial information. The gathered data will be compressed at high speed to enable instant transmission.

“When disaster strikes, it is critical not only to provide available situational awareness as fast as possible but also to rapidly task satellites to change their planned data collection schedules to ensure updated situational information can be acquired and delivered to crisis responders on the timescales they require,” said Simonetta Cheli, ESA Director of Earth Observation.  

ESA will implement the project with the help of three technology companies across Europe: Spanish technology company Deimos, Skylabs, a satellite computer company from Slovenia, and GINA, an emergency response software-as-a-service provider from the Czech Republic.

AI usage to respond to natural disasters has been on the rise. In July, researchers at the University of Southern California in Los Angeles created an AI model that precisely predicts wildfire spread by integrating generative AI with satellite data.

Google also has been working with wildfire authorities to release fire-detecting satellites named FireSat in early 2025.

In 2023, NASA and IBM built a geospatial foundation model, which they expect will speed up scientific research on climate change and allow for the tracking and prediction of floods and wildfires.