Pokémon Go Players Unknowingly Trained AI for Delivery Robots: Niantic Partners with Coco Robotics
Key Takeaways
- ▸Niantic's Visual Positioning System, trained on 30+ billion images from Pokémon Go players, will power autonomous delivery robots through Coco Robotics partnership
- ▸Pokémon Go's crowdsourced data collection—particularly through the 2020 Field Research feature—created detailed 3D models of real-world locations across varying weather, lighting, and angles
- ▸VPS offers centimeter-level location accuracy using visual landmarks, overcoming GPS limitations in urban environments where autonomous delivery currently struggles
Summary
Niantic Spatial, the team behind Pokémon Go, has announced a partnership with Coco Robotics to power autonomous food delivery robots using the company's Visual Positioning System (VPS). The VPS was trained on over 30 billion images captured by Pokémon Go players over nearly a decade, enabling the system to pinpoint locations with centimeter-level accuracy by analyzing buildings and landmarks rather than relying on GPS. Players who spent years hunting virtual creatures unknowingly contributed mapping data through gameplay and especially through the 2020 "Field Research" feature that rewarded players for scanning real-world landmarks and statues. The technology addresses a critical need in autonomous delivery, as GPS often fails in urban environments where tall buildings obstruct satellite signals, causing delivery robots to become lost or delayed. Niantic CEO John Hanke noted that the challenge of rendering Pikachu realistically in augmented reality is fundamentally the same problem as enabling robots to safely navigate city streets—both require understanding real-world spatial relationships through visual analysis.
- The partnership exemplifies how data collected for one purpose can be repurposed years later for unrelated applications, raising questions about user consent and data reuse
Editorial Opinion
This partnership highlights both the practical potential and ethical complexity of crowdsourced data. While repurposing Pokémon Go imagery to improve delivery robot accuracy is genuinely innovative, it underscores how users often have little visibility into how their location data and images are ultimately leveraged. The fact that hundreds of millions of players unknowingly trained critical AI infrastructure deserves scrutiny around transparency and consent, even as the technology demonstrates real benefits for logistics and urban mobility.


