Highlights
- Location scans submitted by Pokémon Go players were used to train AI models developed by Niantic Spatial
- The technology is designed to help drones navigate when GPS signals are unavailable or disrupted
- Niantic Spatial has partnered with drone software company Vantor, which works with military clients
- Privacy advocates have raised concerns about civilian gaming data contributing to defence-related technologies
From augmented reality game to AI training tool
What began as a popular mobile game built around catching virtual creatures in real-world locations has become part of a much broader technological story.
Niantic, the company behind Pokémon Go, collected location scans submitted by players through an optional feature introduced in 2021. The system encouraged users to record real-world landmarks and PokéStops using their smartphone cameras in exchange for in-game rewards.
According to reports, these historical scans were later used to train artificial intelligence models capable of recognising and interpreting physical environments. The models are now owned by Niantic Spatial, a company created after Niantic spun off parts of its business and sold its gaming division in 2025.
AI designed for environments where GPS cannot be trusted
The issue has attracted attention following Niantic Spatial's partnership with Vantor, a company specialising in spatial navigation software for drones, including systems used by some military organisations.
The collaboration focuses on helping autonomous systems operate in areas where GPS signals are unavailable, jammed, spoofed or otherwise unreliable. The technology is intended to allow drones and field teams to maintain accurate positioning and situational awareness without depending entirely on satellite navigation.
Vantor has said future battlefields are likely to involve a wide range of connected systems that need to be integrated quickly, making adaptable navigation software increasingly important.
Both companies have stressed that Pokémon Go scans were not directly handed over to Vantor. However, they confirmed that the player-generated scans were used to train Niantic Spatial's foundation models, which underpin the technology being developed.
A spokesperson for Niantic Spatial said the scans had been submitted voluntarily by players who opted into the feature and were governed by the platform's terms of service and privacy policies at the time.
Privacy questions emerge over secondary use of data
The development has reignited debate over how user-generated data can be repurposed long after it is collected.
Tom Sulston, head of policy at Digital Rights Watch, questioned whether users fully understood the potential future uses of information gathered through a video game. He argued that regulators should focus on whether data practices are fair and in the best interests of users, rather than relying solely on lengthy terms and conditions.
Researchers also note that the use of consumer data in security and defence contexts is not unprecedented. Dr Rob Nicholls, associated with the University of Sydney's Centre for AI, Trust and Governance, pointed to previous instances where fitness-tracking data had revealed the locations of military facilities.
The partnership between Niantic Spatial and Vantor remains in its early stages, but it highlights how information originally gathered for entertainment purposes can later become valuable in entirely different sectors.
The growing overlap between consumer technology, artificial intelligence and defence applications is likely to intensify scrutiny of how companies collect, store and ultimately deploy user-generated data.
