NASA and IBM have launched a new open-source artificial intelligence model that could help scientists study the Moon faster and in greater detail. The model is designed to analyse decades of lunar observation data and could support future plans for a sustained human presence on the Moon.
Called the NASA-IBM Lunar Foundation Model, the AI tool has been trained using more than 30 layers of data collected by nine instruments across four NASA missions. This includes data from NASA’s Lunar Reconnaissance Orbiter, which has been studying the Moon for years.
According to Reuters, the new model can help researchers identify possible ice deposits in areas of the Moon that remain permanently in darkness. It can also be used to map craters, identify potential landing locations and study volcanic features on the lunar surface.
These tasks have traditionally required scientists to manually analyse large amounts of maps and images. The new AI model could make this process faster while also providing more detailed analysis than some existing machine-learning systems.
NASA and IBM said the model performed up to 23 per cent more accurately than commonly used methods in benchmark tests when identifying important features on the lunar surface.
Finding ice on the Moon is particularly important for future space missions. Lunar ice could provide access to water and oxygen, which could eventually be used to support astronauts and potentially produce rocket fuel for longer missions, including future journeys to Mars.
The Lunar Foundation Model is part of IBM and NASA’s Prithvi family of open foundation models, which also includes AI models focused on areas such as weather and geospatial data.
NASA’s Artemis programme aims to send astronauts back to the Moon in 2028. The programme is expected to test technologies and systems needed for longer-term lunar missions and eventually support human exploration of Mars.
