Google has unveiled the Genie 2 artificial intelligence model, a significant advancement from its predecessor that is capable of generating three-dimensional virtual worlds. This new model allows users to navigate these virtual spaces in both first and third person perspectives, showcasing the evolution of AI technology.
The applications of Genie 2 go beyond mere entertainment, as it can also be utilized for educational purposes in training other AI models. By simulating realistic yet controlled environments, AI models can develop specific skills that can be applied in various fields, such as robotics. Google believes that this technology is a stepping stone towards the creation of universal artificial intelligence capable of solving diverse problems.
Similar to its predecessor, Genie 2 operates by taking text or image inputs and generating animations with responsive elements to user commands. While Google has not revealed the specifics of the learning process, it is known that an extensive dataset exceeding 30,000 hours of video material was utilized in the development of the first version.
During demonstrations, Genie 2 showcases interactive scenes featuring a range of characters, from wooden puppets to robots, all of which respond to user commands without the need for a pre-existing game engine. However, the practical applications of these technologies remain uncertain.
A notable feature of Genie 2 is its ability to remember and restore objects that temporarily disappear from view, a task that has posed challenges for other generative models. While the model’s memory is limited to temporary frames, it demonstrates improved stability in maintaining the integrity of the virtual world compared to other AI developments. This advancement marks a significant milestone in the field of artificial intelligence and opens up new possibilities for future applications.