News on Medial

Google introduces image-to-game AI with Genie

Economic TimesEconomic Times · 8m
Google introduces image-to-game AI with Genie

Google DeepMind has unveiled Genie, an AI model that can transform images or sketches into interactive, playable video games. Trained on 200,000 hours of unsupervised internet gaming videos, Genie can generate video games with just a single prompt or image. The model, known as Generative Interactive Environments (Genie), can create action-controllable virtual worlds based on text, images, photos, and hand-drawn sketches. While not publicly available, Genie's foundation model can identify main characters and allow users to control them within the generated world. This development presents significant potential for game developers looking to rapidly prototype ideas. However, Genie currently faces challenges such as slow playback and competition from established game development tools.

Comments

Download the medial app to read full posts, comements and news.