Pokémon GO Data May Have Helped the Military Train Drones 

In 2016, Niantic released a mobile game that “made” players walk the streets looking for Pokémon, which they could catch through their camera lens, and the whole world went crazy for it, reaching more than 800 million downloads by 2018. 

What none of the millions of players thought about at the time was that they were providing data that would be used to train AI models, which could eventually prove useful for military drones. 

Pokemon trainers contributed to this without knowing it by agreeing to terms whose future implications were difficult to foresee. 

How did searching for Pikachu become Drone training?

The biggest red flag came in 2021 with a new PokéStop feature. To earn in-game rewards, trainers had to record real-world locations and upload the footage through the app. Those terms allowed Niantic to use about 30 billion scans of streets, parks, buildings that were gathered from players across the globe.

According to Dutch publication Trouw, Niantic used this footage to build large-scale 3D maps of real-world environments and train its Visual Positioning System (VPS). A system like this helps devices figure out where they are based on what their cameras see instead of relying only on GPS. It was also used to train AI models to recognize and understand physical spaces.  

This probably wouldn’t raise any motion if in 2025 Niantic Spatial, a spin-off company from Niantic, had not announced its partnership with Vantor, a company that specializes in spatial detection software for drones, including ones used by some militaries. 

What did Vantor get?

The goal of this partnership was unified air-to-ground positioning for military operations in GPS-denied environments. Which is what modern warfare relies on increasingly. 

Vantor’s chief product officer, Peter Wilczynski, said in a December interview: “The partnership addresses a critical vulnerability in modern operations: GPS unavailability, spoofing, interference, and jamming.”

Vantor gained access to advanced spatial computing expertise focused on how machines interpret real-world environments in 3D. 

But both companies told The Guardian Australia that ground scans from the game were not given to  Vantor as part of the partnership, but that Pokémon GO scan data was used to train Niantic’s foundation models. 

As stated by the Niantic spokesperson: “AR Scans collected through Pokémon Go were submitted voluntarily by players who opted into the feature and were subject to the applicable Terms of Service and Privacy Policy at the time.” 

In February, Vantor was awarded a US Army contract worth up to $217 million (£162m) for training software focused on building realistic simulation environments for military use.

GPS jamming and spoofing where a counterfeit radio signal is emitted to suggest a false location are used in a host of conflicts, including the Ukraine and Iran wars. They are designed to mislead kamikaze and reconnaissance drones as well as GPS guided missiles.

The visual positioning system, an alternative to GPS, allows for determining location using a camera that matches what it observes against a pre existing detailed 3D model of the environment. It is resistant to traditional signal jammers.

Can your location data be used in unexpected ways? 

For travelers, this debate extends beyond gaming. Modern services such as navigation apps, ride-hailing platforms, social media check-ins, and even eSIM-powered connectivity rely on location and network data to function. Most of that information is used for legitimate purposes such as improving coverage and services.

The Pokémon GO story is a reminder that data collected for one purpose can sometimes find value in entirely different applications years later. As digital services become more deeply embedded in everyday life, understanding how data is collected and used may become more important than the service itself.