Google DeepMind team has introduced Genie, an advanced AI model capable of creating interactive 2D video games from mere text or image prompts.

Genie represents a pioneering venture by Google to focus on the video game industry, the world’s highest-grossing entertainment sector. Similar to Sora, drawing on a “world model” concept, this AI has been trained on over 200,000 hours of unlabelled video footage from 2D platformer games available on the internet. This training enables Genie to understand and replicate the complex dynamics of video games, creating new ones from a single prompt or image.

The process behind Genie’s magic involves complex AI models that can understand and simulate physics, game mechanics, and player interactions without human supervision. Its creators have designed it as an “actionable-controllable world model,” distinct from any previous AI efforts in the gaming space. Google DeepMind’s Genie can take sketches, photographs, and other images, including those never seen before, to create interactive environments that users can explore and manipulate.

Genie’s unsupervised learning from internet videos make it learn and reproduce controls for in-game characters without any manual labeling. With a colossal 11 billion parameters, it features a spatiotemporal video tokenizer, an autoregressive dynamics model, and a scalable latent action model. These components work together seamlessly to bring virtual worlds to life, frame by frame.

Despite its impressive capabilities, Genie currently operates with certain limitations, such as generating games at a low frame rate of 1FPS, and remains a research project within Google DeepMind, not yet available to the general public.

Beyond the potential to democratize game development, allowing individuals without programming expertise to bring their visions to life, Genie also holds promise for applications in robotics and other areas requiring simulation of real-world environments.

However, this innovation also prompts questions regarding the future of game development jobs and the potential for misuse. Google has acknowledged these concerns, emphasizing the importance of ethical considerations in Genie’s development and application.