How Pokémon Go Players Are Unknowingly Training AI for Real-World Navigation
- Playing Pokémon Go does more than entertain; it contributes to data collection for a 3-D world map.
- Niantic's recent blog post highlighted its “large geospatial model” technology.
- Players generate data through the visual positioning system (VPS) used in Pokémon Go’s features, like Pokemon Playgrounds.
Pokémon Go Players Help Train AI Models
Playing Pokémon Go does more than entertain; it contributes to data collection for a 3-D world map. Developer Niantic uses this data to improve artificial intelligence (AI) navigation.
Niantic’s recent blog post highlighted its “large geospatial model” technology. This technology helps machines perceive and interact with physical spaces. It aims to assist with augmented reality (AR) applications, robotics, content creation, and autonomous systems.
Players generate data through the visual positioning system (VPS) used in Pokémon Go’s features, like Pokemon Playgrounds. Niantic claims it collects “one million fresh scans each week” to create a detailed understanding of the environment.
Niantic’s goal is to build spatial intelligence for AR gaming, logistics, and design applications. This vision was shared in 2018 and is outlined in Niantic’s privacy policy. Some players feel responsible for the data collection due to their gaming habits, while others see it as a normal business practice.
Niantic insists using the scanning feature is optional. Players need to visit specific locations to contribute data. Simply playing the game does not train the AI model.
Concerns about data collection are rising. Anton Dahbura from Johns Hopkins University warns that location data is crucial for national security. Misuse of this information could have serious implications.
While some players view this data harvesting as old news, the potential risks should not be ignored. Understanding how location data can affect various sectors, including transportation and food manufacturing, is essential.