Google DeepMind has introduced a new version of its artificial intelligence (AI) tool called Genie 3, which creates 3D digital spaces that can be explored in real time.
These environments respond to simple text prompts and allow both people and AI systems to move through them, similar to how video game worlds work.
This approach is based on a “world model”, a type of system designed to simulate environments for learning, entertainment, or training purposes. Unlike traditional video games, where artists build the world by hand, Genie 3 generates everything using AI.
Did you know?
Subscribe – We publish new crypto explainer videos every week!
What is Polkadot in Crypto? (DOT Animated Explainer)
According to Google, Genie 3 can support a few minutes of interaction with steady visuals. If something is placed in the environment, such as writing on a chalkboard or paint on a wall, it will remain visible for up to a minute, even when not being observed.
Scenes are displayed at 720p resolution and run at 24 frames per second, which makes the experience smoother.
Another new feature allows users to add things to the environment while it is running. By typing a new prompt, it is possible to change the weather or introduce new characters without starting over. These updates happen in real time, which helps the environment feel more responsive.
Despite the improvements, access to Genie 3 is still limited. It is being released as a research tool, not for general public use.
On July 30, Google tested AI in the US that estimates user age based on account activity like searches and YouTube views. How? Read the full story.
