This article was published on March 26, 2024

ESA to build ChatGPT-style Earth observation digital assistant

The chatbot will translate troves of satellite data into actionable insights


ESA to build ChatGPT-style Earth observation digital assistant

The European Space Agency (ESA) has announced it will build a ChatGPT-style digital assistant to help humans better decode complex Earth observation data.

Led by Φ-lab (Phi-Lab) — the agency’s Earth observation innovation and investment arm — the chatbot will be trained to understand human queries about the Earth and respond with human-like answers. The lab is due to kick off the project next month.

Earth observation tools like satellites and weather stations gather vast volumes of data on our home planet every day. Most of this information is unlabelled, which makes it hard for traditional machine learning algorithms to make sense of what they are seeing. 

That is why ESA Φ-lab is working on developing a foundational AI model to underpin the digital assistant. Foundational models, like those used to power ChatGPT, are AI neural networks trained on unlabelled datasets and designed to handle a wide variety of tasks. 

This means ESA’s chatbot will be able to take unlabelled Earth observation data straight from satellites and quickly translate it into meaningful insights. 

“The concept of an Earth observation digital assistant that can provide a broad range of insight from varied sources is a tantalising prospect,” said Giuseppe Borghi, head of ESA Φ-lab.

The lab has several ongoing initiatives for creating foundation models dedicated to Earth observation. These models provide critical information on topics such as how to minimise damage from extreme weather events or the amount of methane pollution in the atmosphere. 

Richat_structure_ESA
The 40 km-diameter circular Richat structure in the Sahara desert is one of the many geological features the ESA is using to test its AI models. Credit: ESA

One foundation model, PhilEO, launched at the beginning of 2023, and is now reaching maturity. The model can recognise features like the Richat Structure above without human supervision or labelling.

“Given the extremely encouraging progress already achieved with PhilEO, I fully expect the new projects to yield game-changing results in the near future,” said Borghi.

The project comes amid a surge of new, task-specific chatbots, aimed at helping people understand everything from climate change to their finances.

Get the TNW newsletter

Get the most important tech news in your inbox each week.

Also tagged with