How Pokémon Go Helped Build an AI Model by Collecting Your Data

Summary:
Niantic, the company behind Pokémon Go, has been using players’ data to create a “Large Geospatial Model” (LGM) to improve robot and device navigation. By collecting over 10 million scanned locations, Niantic trains neural networks to understand and interact with physical spaces, aiding advancements in AR, robotics, and autonomous systems. The data comes from in-game scans made by players, contributing to the model’s development. While many users are unaware, critics worry that this technology could eventually be used in military applications, such as automated weapons.

Since its launch in 2016, *Pokémon Go* has captured the attention of millions, but many players didn’t realize they were unknowingly helping to build a powerful artificial intelligence (AI) model. Niantic, the company behind the game, has been using data from users’ in-game scans to create a “Large Geospatial Model” (LGM) aimed at improving navigation for robots and other devices. This model allows machines to better understand and interact with the physical world.

Niantic’s data collection has resulted in over 10 million scanned locations, with players contributing an additional 1 million scans each week. This information has been used to train neural networks to operate in millions of locations worldwide, forming a global understanding of geographic spaces. The LGM helps computers not only perceive physical environments but also interact with them, contributing to advancements in augmented reality (AR) glasses, robotics, autonomous systems, and more.

Just as AI models like ChatGPT rely on large datasets of text, Niantic’s LGM uses geodata to help machines infer the appearance of physical locations. For humans, this spatial awareness is natural, but for machines, it’s a complex task. Niantic’s Visual Positioning System, which is based on smartphone camera images, enables precise location identification down to a centimeter.

While many *Pokémon Go* players are unfazed by the data collection, critics express concerns about the broader implications. Some fear that this technology could eventually be used in military applications, such as automated weapon systems.

Stay Updated!
Join our WhatsApp Channel for the latest updates, exclusive content, and more! Click the link below to join now:
👉 Join Our WhatsApp Channel