×
The first fully AI generated video game is getting rave reviews
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

AI’s first fully-generated video game demonstrates both the potential and current limitations of generative AI in game development, marking a significant milestone in the intersection of artificial intelligence and gaming.

The breakthrough: Oasis, a Minecraft-inspired game released by Israeli AI startup Decart and silicon design company Etched, represents the first fully AI-generated video game where each frame is created in real-time by an AI model.

  • The game uses a transformer AI model trained on Minecraft gameplay footage to generate each new frame based on the previous one and user inputs
  • Unlike traditional video games that rely on game engines and hand-coded rules, Oasis creates its entire visual experience through AI generation
  • Players can access the game for free online, offering a unique opportunity to experience AI-generated gameplay firsthand

Technical implementation: The game leverages transformer-based AI algorithms running on specialized hardware to create a dynamic, ever-changing gaming environment.

  • The AI model functions similarly to large language models but is specifically trained on Minecraft gameplay data
  • Custom silicon optimization by Etched enables the real-time generation of game frames
  • Players can even upload their own images for the AI to incorporate into the game environment

Unique characteristics: The AI-generated nature of Oasis creates an unstable and surreal gaming experience that sets it apart from traditional video games.

  • The game world constantly shifts and morphs as the AI generates new frames
  • Players have discovered they can “trick” the AI into creating new environments, including dark moonscapes
  • The game features unusual artifacts like misshapen animals and nonsensical architectural elements

Expert perspectives: Gaming and AI specialists offer mixed views on Oasis’s significance and practical applications.

  • NYU Professor Julian Togelius notes that while impressive, the technology is currently too inconsistent for conventional gaming applications
  • Game designer Frank Lantz describes Oasis as residing in an “uncanny valley” that impacts its playability, while acknowledging its innovative nature
  • The game’s viral success stems largely from players experimenting with ways to manipulate the AI system

Future implications: While Oasis may not represent a immediately viable approach for mainstream game development, it suggests interesting possibilities for the future of AI in gaming.

  • The technology could potentially be applied to character control and world generation in more traditional game formats
  • The project demonstrates both the creative potential and current limitations of AI in game development
  • The success of this experiment may inspire future innovations in AI-driven game design

Innovation versus practicality: While Oasis represents a fascinating technical achievement, its primary value may lie in demonstrating the possibilities and challenges of AI-generated gaming experiences rather than establishing a new standard for game development.

The First Entirely AI-Generated Video Game Is Insanely Weird and Fun

Recent News

Radiologists thrive despite predictions of their demise through AI replacement

AI tools enhance radiologist productivity by automating routine tasks and providing diagnostic assistance, while professional expertise remains essential for medical image interpretation.

Privacy advocacy group NOYB challenges Meta’s use of European data for AI training

European privacy watchdog challenges Meta's use of "legitimate interest" justification for AI training on user data, threatening collective action that could result in massive financial penalties.

Neural networks trained on quantum physics are changing molecular science

Researchers combine decentralized computing networks with atomic-level simulations to predict drug interactions at speeds up to a million times faster than traditional methods.