true77
Generative AI developers at Google DeepMind have introduced “Genie,” a new artificial intelligence system that can create playable worlds resembling video games from images and text prompts. Genie, announced in a recent paper, is not yet available to the public.
Training on over 200,000 hours of publicly available 2D platformers, Genie can take prompts, sketches, and images to generate virtual worlds for users to play in, creating assets from scratch based on player actions. The AI’s impressive understanding of physics through unsupervised training allows it to grasp different game mechanics, potentially extending its use to training robots on navigation.
While not as high-quality as Sora, Genie raises questions about the future job security of video game developers and the impact on game marketplaces like Steam. Google Bard has been rebranded as Google Gemini, and discussions around AI in gaming continue.
Topics:
– Gaming
– Google
