Archive — US — techcrunch
Could Google’s Veo 3 be the start of playable world models?
Demis Hassabis, CEO of DeepMind, hinted that Google’s Veo 3 video generation model could lead to playable world models for video games. While currently a passive tool, Veo 3 may evolve with Gemini 2.5 Pro to create interactive environments. Google is positioning itself amidst competition from AI firms in game development.
Key points
- Demis Hassabis teased the potential of Veo 3 for gaming in response to user interest.
- Veo 3 can generate videos but is not yet an interactive world model.
- Gemini 2.5 Pro aims to simulate human brain aspects for dynamic modeling.
- DeepMind previously introduced Genie 2 for creating endless playable worlds.
- Competitors like Fei-Fei Li's World Labs are also developing game-like 3D scene generation.
Why it matters
- In 2025, AI and machine learning trends are pushing for real-time simulations and interactive experiences.
- Past developments in generative AI are now intersecting with gaming, emphasizing AI's influence on entertainment.
- Google's strategic investments in AI technology reflect broader industry movements towards hybrid models that combine generative and interactive capabilities.
Google's exploration into playable world models via Veo 3 signifies a pivotal moment in AI and gaming, with the company poised to shape future interactive entertainment while navigating a competitive tech landscape.