{"id":1760,"date":"2024-08-31T18:46:21","date_gmt":"2024-08-31T18:46:21","guid":{"rendered":"https:\/\/www.inthacity.com\/blog\/?p=1760"},"modified":"2025-07-06T19:46:12","modified_gmt":"2025-07-07T00:46:12","slug":"ai-generative-doom-game-creation-future","status":"publish","type":"post","link":"https:\/\/www.inthacity.com\/blog\/tech\/ai-generative-doom-game-creation-future\/","title":{"rendered":"Doom and AI: How Google\u2019s Neural Network Created a Playable Game Without Code (and What That Means for the Future of Gaming)"},"content":{"rendered":"<p>In the realm of classic video games, few titles hold the iconic status of <em>Doom<\/em>. First released in 1993, this pioneering first-person shooter has seen its code run on everything from toasters to calculators, cementing its place in both gaming and pop culture. But now, <em>Doom<\/em> has been resurrected in a way that even its original creators couldn't have imagined: through the magic of generative AI. Yes, you read that right\u2014Google\u2019s AI has managed to recreate the game without a single line of original code. Welcome to the future of gaming, where your next favorite game might just be \u201challucinated\u201d by a neural network.<\/p>\n<h4>Doom\u2019s Resurrection by AI: A New Kind of Gaming Magic<\/h4>\n<p>Let\u2019s start with the basics. A team at <a rel=\"noopener\" target=\"_new\" href=\"https:\/\/research.google\/\">Google Research<\/a> led by Dani Valevski has developed a model named GameNGen that can generate a playable version of <em>Doom<\/em>\u2014not by copying its code, but by \u201cwatching\u201d it being played. Imagine watching your favorite show so many times that you could recreate it from memory, scene by scene. That\u2019s essentially what GameNGen has done, only with more algorithms and less popcorn.<\/p>\n<p>According to a report from <a rel=\"noopener\" target=\"_new\" href=\"https:\/\/www.newscientist.com\/\">New Scientist<\/a>, this AI marvel was trained on 300 million pieces of gameplay data, observing every cough, splatter, and demon roar. The result? A model that can generate a 20-second clip of <em>Doom<\/em> gameplay that\u2019s nearly indistinguishable from the real thing. Players can shoot enemies, open doors, and navigate the hellish landscapes just as they would in the original game. But here\u2019s the kicker\u2014there\u2019s no original game code or graphics involved.<\/p>\n<h4>How Does It Work? Neural Networks and Hallucinations<\/h4>\n<p>The process behind this digital sorcery is as fascinating as it is complex. The researchers first trained an AI model to play <em>Doom<\/em> like a human. This wasn\u2019t your typical AI opponent; instead, it was more like a method actor, fully immersing itself in the role of a <em>Doom<\/em> marine. Meanwhile, a second model, based on the Stable Diffusion image generator, learned how the game\u2019s environment changed in response to player actions. Essentially, it became a virtual mimic, understanding the rules of <em>Doom<\/em> without ever seeing its code.<\/p>\n<p>The final product is a kind of \u201challucination\u201d\u2014a game environment and interaction system conjured by AI that replicates the experience of <em>Doom<\/em> without actually being <em>Doom<\/em>. It\u2019s like playing a game created by someone who\u2019s only ever heard about <em>Doom<\/em> in great detail. And here\u2019s the wild part: in tests, human players could barely tell the difference between the AI-generated clips and the real game.<\/p>\n<p style=\"text-align: center;\"><img  title=\"\" loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-1763 size-large\" src=\"https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/Doom-Eternal-Game-1024x576.jpg\"  alt=\"Doom-Eternal-Game-1024x576 Doom and AI: How Google\u2019s Neural Network Created a Playable Game Without Code (and What That Means for the Future of Gaming)\"  width=\"640\" height=\"360\" srcset=\"https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/Doom-Eternal-Game-1024x576.jpg 1024w, https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/Doom-Eternal-Game-300x169.jpg 300w, https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/Doom-Eternal-Game-768x432.jpg 768w, https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/Doom-Eternal-Game-1536x864.jpg 1536w, https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/Doom-Eternal-Game-600x338.jpg 600w, https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/Doom-Eternal-Game.jpg 1920w\" sizes=\"auto, (max-width: 640px) 100vw, 640px\" \/><\/p>\n<h4>The Implications: Is AI About to Replace Game Developers?<\/h4>\n<p>Naturally, this raises some big questions. Are we on the cusp of a revolution where AI replaces human game developers? Will future games be spun from the minds of <a class=\"wpil_keyword_link\" href=\"https:\/\/www.inthacity.com\/blog\/tech\/neural-networks-ai-revolution-how-they-work-why-they-matter\/\"   title=\"neural networks\" data-wpil-keyword-link=\"linked\"  data-wpil-monitor-id=\"203\">neural networks<\/a>, bypassing the need for human creativity altogether? According to Andrew Rogoyski at the <a rel=\"noopener\" target=\"_new\" href=\"https:\/\/www.surrey.ac.uk\/\">University of Surrey<\/a>, UK, the answer is a resounding \u201cnot yet.\u201d<\/p>\n<p>Rogoyski points out that while AI can mimic the mechanics of a game, it\u2019s not about to replace the human element that makes games truly engaging. \u201cI don\u2019t think it\u2019s the end of those game studios,\u201d he says. \u201cWhat the game studios have is the imagination, the skills, to actually create these worlds, to understand gameplay, to understand engagement, to understand how to draw us into a story. It\u2019s not just the nuts and bolts, the bits and bytes.\u201d<\/p>\n<p>In other words, AI can recreate the <em>Doom<\/em> experience, but it can\u2019t capture the essence of what makes a game like <em>Doom<\/em> an enduring classic. The thrill of exploration, the carefully crafted level designs, the pacing\u2014these are things that, for now, require a human touch.<\/p>\n<h4>What\u2019s Next? The Future of AI in Game Development<\/h4>\n<p>But don\u2019t count AI out just yet. The researchers behind GameNGen see this as a proof-of-concept that could revolutionize how games are made in the future. Imagine generating a game from a simple text description or concept art, drastically reducing the time and cost involved in game development. No more crunch time, no more burning the midnight oil\u2014just feed the AI your ideas and watch it generate the game of your dreams.<\/p>\n<p>Of course, this also opens the door to some pretty wild possibilities. What if AI starts creating entirely new genres of games, ones that we humans haven\u2019t even thought of yet? The creativity of AI is still a great unknown, and as it continues to evolve, we might find ourselves playing games that are as much a surprise to the developers as they are to us.<\/p>\n<h4>Final Thoughts: A Brave New World of Gaming<\/h4>\n<p>So, should game developers start updating their resumes? Not just yet. While GameNGen is an impressive achievement, it\u2019s not about to replace the human element in game design. But it does hint at a future where AI plays a much bigger role in the creative process, potentially leading to a new era of game development where human imagination and AI innovation go hand in hand.<\/p>\n<p>In the meantime, if you ever find yourself playing a game that feels a bit too perfect\u2014or a bit too weird\u2014you might want to check if it was made by a neural network. Because in the world of AI, the line between reality and hallucination is getting blurrier by the day.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Generative AI has now recreated the classic Doom game\u2014without a single line of original code. Could this be the future of game development?<\/p>\n","protected":false},"author":2,"featured_media":1761,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[270,21],"tags":[318,321],"class_list":["post-1760","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","category-tech","tag-generative-ai","tag-neural-networks"],"aioseo_notices":[],"jetpack_featured_media_url":"https:\/\/www.inthacity.com\/blog\/wp-content\/uploads\/2024\/08\/DALL\u00b7E-2024-08-31-14.43.26-A-stunning-16_9-feature-image-for-an-article-titled-Doom-and-AI_-How-Googles-Neural-Network-Created-a-Playable-Game-Without-Code.-The-image-should-.webp","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/posts\/1760","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/comments?post=1760"}],"version-history":[{"count":0,"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/posts\/1760\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/media\/1761"}],"wp:attachment":[{"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/media?parent=1760"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/categories?post=1760"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.inthacity.com\/blog\/wp-json\/wp\/v2\/tags?post=1760"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}