Pokémon Go images train Niantic Spatial's geospatial AI

Niantic uses images captured through the Pokémon Go app to train Niantic Spatial's geospatial AI model. The model, built on 30 billion urban images, enables centimeter-level location precision for applications like delivery robots. Niantic's terms disclose that such images are banked as mapping data.

Niantic, developer of Pokémon Go, is leveraging player-captured images to advance Niantic Spatial's work in geospatial AI. Niantic Spatial aims to create digital maps allowing machines, robots, and AR glasses to navigate the physical world with centimeter-level precision, according to an MIT Technology Review article cited by Eurogamer. The AI model has been trained on 30 billion images from urban environments, often taken near in-game hotspots like gyms from various angles and times of day, as reported by both Eurogamer and Kotaku. This data supports a visual positioning system that surpasses traditional GPS in signal-interfered city areas. Niantic Spatial collaborates with Coco Robotics, a startup deploying last-mile delivery robots in US and Europe cities. The partnership originated when AR was seen as the future, though its applications have evolved. Niantic Spatial CTO Brian McClendon stated, “Maps are not only becoming more detailed; they are being used more and more by machines.” He added, “We had a million-plus locations around the world where we can locate you precisely... We know where you’re standing within several centimeters of accuracy and, most importantly, where you’re looking.” Niantic Spatial CEO John Hanke described the Coco Robotics partnership as the start of a vision for a dynamic virtual world simulation that updates with real-world changes and incorporates data from more robots. Eurogamer describes Niantic Spatial as a spinout from Niantic, while Kotaku calls it an AI-focused offshoot formed last year following Niantic's acquisition by Saudi-owned Scopely. Niantic's terms and conditions note that images are banked as mapping data.

Relaterede artikler

Illustration depicting Google's Nano Banana 2 AI image generator launch, featuring the Gemini app interface generating realistic images with speed and precision.
Billede genereret af AI

Google releases Nano Banana 2 AI image generator

Rapporteret af AI Billede genereret af AI

Google has launched Nano Banana 2, an updated AI image generation model that combines the precision of its Pro version with faster processing speeds. Powered by Gemini 3.1 Flash Image, the tool becomes the new default across Google's platforms, including the Gemini app and search features. It aims to deliver more accurate and realistic images while expanding creative capabilities.

Experts argue that physical AI, involving robots and autonomous machines interacting with the real world, may provide a direct path to artificial general intelligence. Elon Musk's comments on Tesla's Optimus robots highlight this potential, amid growing investments in related technologies. The year 2026 is seen as a key inflection point for the field.

Rapporteret af AI

Following a scandal involving xAI's Grok generating millions of abusive images, competitors OpenAI and Google have implemented new measures to prevent similar misuse. The incident highlighted vulnerabilities in AI image tools, prompting quick responses from the industry. These steps aim to protect users from nonconsensual intimate imagery.

In the new life simulation game Pokémon Pokopia, players are using laser sensors to create automated systems for town maintenance. These devices, crafted from PokeMetal, enable motion-activated mechanisms that handle tasks like watering crops without constant player input. Early designs include logic gates and water flow channels shared on platforms like Reddit and Twitter.

Dette websted bruger cookies

Vi bruger cookies til analyse for at forbedre vores side. Læs vores privatlivspolitik for mere information.
Afvis