Pokémon Go data helped train AI now linked to military drones

2026-06-15 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Fundamental Awareness, quick

Summary

Niantic Spatial's AI models, initially trained using millions of voluntary 3D scans from Pokémon Go players, are now central to a partnership with US defense contractor Vantor. These scans, collected since a 2021 in-game update, generated billions of visual mapping data points, enabling camera-based navigation without GPS. The collaboration, announced in December 2025, aims to provide GPS-denied navigation for military drones and autonomous systems, combining Niantic's Visual Positioning System with Vantor's Raptor software and 3D terrain data. Early tests showed up to 70 percent error reduction and 1.5 meters accuracy. Vantor also secured a US Army contract worth up to \$217 million in February 2026 for 3D terrain data for the One World Terrain program. Niantic's gaming division was acquired by Saudi-backed Scopely for \$3.5 billion in March 2025, with Niantic Spatial continuing as a standalone spatial AI company.

Key takeaway

For AI ethicists and policy makers evaluating data privacy and military applications, this case highlights the unforeseen dual-use potential of consumer-generated data. Your understanding of data governance must evolve to address how seemingly innocuous user contributions can train foundational models later integrated into defense systems. Consider advocating for clearer consent mechanisms and transparency regarding potential downstream uses of crowdsourced data.

Key insights

Volunteer user data, even from games, can train foundational AI models with significant defense applications.

Principles

• Spatial AI enables GPS-free navigation.
• Diverse data sources enhance model robustness.

Method

Niantic's Visual Positioning System combines with Vantor's Raptor software and satellite-derived 3D terrain data to create a shared coordinate system for GPS-denied operations.

In practice

• Develop AI for GPS-denied environments.
• Utilize crowdsourced data for foundational models.

Topics

• Niantic Spatial
• Vantor
• GPS-denied Navigation
• Spatial AI
• Dual-use Technology
• Crowdsourced Data
• Military Drones

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Tech Journalist, Policy Maker, AI Ethicist

Related on AIssential

• Pokémon Go players unwittingly contributed to tech with military drone uses — Player-generated geospatial data from games like "Pokémon Go" is being repurposed for advanced navigation systems, including military applications, raising ethical questions about consent.
• Pokémon Go data trained AI that could assist military drones in war zones — Pokémon Go user location data trained AI now leveraged for military drone navigation in GPS-denied environments.
• The Download: Pokémon Go to train world models, and the US-China race to find aliens — Crowdsourced AR data can build world models for robot navigation, while geopolitical rivalries impact space exploration and AI ethics.
• Project Maven and the Age of AI Warfare — AI integration in warfare accelerates decision-making and expands surveillance, raising profound ethical and strategic questions.
• Turns Out Niantic Needed Your Pokemon Go Photos To Help Delivery Robots Navigate The World — Crowdsourced geospatial data from AR games can significantly enhance autonomous robot navigation accuracy.
• The Pentagon is planning for AI companies to train on classified data, defense official says — The Pentagon is planning to allow generative AI companies to train military-specific models on classified data within secure environments, a significant shift from current practices where models only…

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.