Millions of players thought they were just catching Pokémon. In reality, they were helping build one of the biggest artificial intelligence datasets ever created. Niantic, the company behind Pokémon Go, has now confirmed that photos and AR scans from its hit mobile game have generated more than 30 billion real-world images. That data, collected from over 143 million users, across parks, sidewalks, storefronts, and landmarks, is now being used to train visual navigation systems for robots and autonomous delivery machines.
Players weren’t simply roaming the streets for fun. Every time they aimed their phones at a PokéStop or gym, they were capturing detailed images of their surroundings—at all hours, under different lighting, through rain, snow, or sunshine. What mapping company could ever afford to gather that kind of data so fast or so comprehensively? None. Yet Niantic pulled it off, pixel by pixel, photo by photo, over eight years of gameplay.
POKÉMON GO PLAYERS TRAINED 30 BILLION IMAGE AI MAP
Niantic says photos and scans collected through Pokémon Go and its AR apps have produced a massive dataset of more than 30 billion real-world images.
The company is now using that data to power visual navigation for delivery… pic.twitter.com/FIs65uO3sx
— NewsForce (@Newsforce) March 15, 2026
Now, what started as entertainment has become something else entirely. The company admits that these AR scans and images power “visual positioning” for delivery robots, allowing them to locate themselves on city streets without GPS. Think about that. A game many kids and families downloaded for fun has quietly built a world-scale AI dataset—one that rivals anything collected by big tech’s driverless car fleets.
The real question is whether players ever realized what they were giving away. Did they know their local park photos would someday guide robots instead of catching a Charizard? Niantic says it’s all part of improving AR technology, but the scope of this data trove raises serious questions about consent, privacy, and the unseen price of play. As it turns out, the most powerful AI systems aren’t built by engineers in labs. They’re built by ordinary people who never knew they were helping write the next chapter of machine intelligence.
***
English (US) · 