Glossary
October 22, 2026

Google's AI Lab Releases a Free Map Showing Which of the World's Trees Are Wild Forest and Which Are Farms

Google DeepMind and two environmental research groups released a free AI-made map that sorts every patch of land on Earth, down to the size of a small house, into wild forest, planted forest, tree farms and other land. Satellite maps have long shown where trees are; this one tries to show what kind they are. It's built to help check that coffee, cocoa, palm oil and wood don't come from cleared forest.

Google DeepMind, Google's AI research lab, released a world map today that tries to answer a question satellite photos usually can't: what kind of forest is this?

From space, an ancient rainforest and a neat row of palm-oil trees can both look like "tree cover." But they're very different. A natural forest that has never been cut stores far more carbon and holds far more kinds of plants and animals than a timber plantation harvested every decade or two. When a wild forest is cleared and replanted as a farm, the area can still look green from above, so the loss can go unnoticed.

The new map, called Forest Typology, or ForTy, sorts every 10-by-10-meter patch of land on Earth into one of six groups: untouched forest, forest growing back naturally, forest planted by people, plantations, tree crops like coffee or cocoa grown among trees, and everything else. An AI program learned to tell them apart from satellite images and about 1.7 million locations that experts had already labeled. Checked against more than 9,000 randomly chosen spots, each reviewed by two experts, it picked the right forest type about 90% of the time, and told forest from non-forest about 95% of the time.

The main map shows the world in 2020, a date chosen on purpose. A new European Union rule requires that goods like coffee, cocoa, palm oil and wood sold in Europe don't come from land that was forest at the end of 2020 and has since been cleared. The team also released early yearly maps for 2017 through 2025 showing where natural forest was replaced by plantations or farms. Both are free for anyone to use. The work was done with the World Resources Institute, an environmental research group, and the International Institute for Applied Systems Analysis, a scientific research institute, and is described in a scientific paper.

The makers are frank about the limits. The yearly maps are an early version they plan to keep improving. Telling forest types apart is hard because it depends partly on how land is used, not just how it looks. Accuracy may vary from place to place, and the map is meant as a first screening, not proof on its own that a farm did or didn't clear forest.

Sources

  1. Google DeepMind — Beyond tree cover: mapping the world's forest typesdeepmind.google · primary
  2. Scientific Data — Global forest typology at 10-meter resolution for forest and land-use monitoring (the paper)nature.com
  3. AlphaSignal — Google DeepMind's ForTy maps every forest type on Earth at 10-meter detailalphasignal.ai
Follow the timeline
Get an email when new entries are added.
© 2026 Sugarpine