{
  "id": 6926125,
  "title": "NASA Uses AI to Read the Moon Before Humans Return to Its Surface",
  "url": "https://urgent.news/2026/09/12/nasa-uses-ai-to-read-the-moon-before-humans-return-to-its-surface",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-12T13:04:37.000Z",
  "source": {
    "name": "Colombia One",
    "slug": "colombia-one",
    "url": "https://colombiaone.com/2026/09/12/nasa-uses-ai-to-read-the-moon/"
  },
  "original_language": "en",
  "account": "NASA's Lunar Reconnaissance Orbiter, a spacecraft in orbit around the Moon, has spent more than 17 years gathering detailed mapping data of our celestial neighbor. This extensive data collection has become a crucial resource for training a new artificial intelligence (AI) system developed by NASA and IBM, called the NASA-IBM Lunar Foundation Model. Launched on September 10, 2026, this model aims to help scientists uncover new insights into the Moon's features, such as craters, potential ice deposits, and geological formations.\n\nThe model was trained using data from the LRO, which has produced a vast amount of information on the Moon's surface. This training dataset consists of 2 million image tiles, including high-resolution 1-meter captures and multispectral frames at 100-meter resolution. Additionally, the model incorporates data from other NASA missions, such as GRAIL, Lunar Prospector, and Japan's SELENE (Kaguya), enhancing its understanding of the Moon's complex geological processes.\n\nOne of the model's key applications is identifying areas on the Moon where potential water ice may be present, particularly near the polar regions. These permanently shadowed regions remain cold enough to preserve ice deposits for billions of years. By estimating where surface and subsurface ice may exist, the model helps pinpoint target areas for further study, which could provide a vital resource for future lunar missions.\n\nThe model also excels in crater mapping, a technique used to determine the surface age of the Moon and reconstruct its geological history. Automating this process saves researchers significant time and effort that can be better spent analyzing the implications of these geological formations for understanding the solar system's evolution. In benchmark tests, the model has demonstrated performance equal to or better than existing systems in detecting and measuring craters.\n\nAnother exciting feature of the model is its ability to identify young volcanic formations known as irregular mare patches. Although the Moon no longer experiences volcanic activity, these relatively young formations offer valuable insights into the Moon's past geological evolution. Detailed mapping of these patches will help scientists refine their models of thermal evolution and establish accurate timelines for lunar volcanic events. This information is crucial for planning robotic and crewed missions to the Moon, guiding them towards the most promising locations for collecting fresh geological samples.\n\nThe open-source nature of the NASA-IBM Lunar Foundation Model is a significant aspect of this project. By releasing the model on Hugging Face and publishing its source code on GitHub, NASA has made it accessible to the global research community. Researchers can now build upon this foundation, adapt the model for new tasks using smaller datasets, and contribute their improvements back to the community. This collaborative approach accelerates scientific discovery and benefits a wide range of stakeholders, including universities, space agencies, and commercial ventures involved in lunar exploration.",
  "summary": "NASA and IBM Research have launched an open-source artificial intelligence model to extract new insights from nearly two decades of lunar data, aiding future crewed missions to the Moon. Announced Sept. 10, 2026, the NASA-IBM Lunar Foundation Model helps scientists map craters, locate potential ice deposits, and identify geological features. Co-developed with academic institutions, it […]",
  "key_points": [
    "NASA Lunar Reconnaissance Orbiter has mapped Moon for over 17 years.",
    "NASA-IBM Lunar Foundation Model uses LRO data to identify ice deposits and craters.",
    "Model's open-source release accelerates lunar exploration research."
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}