A team of researchers from the University of Lyon, Purdue and Ubisoft have published a paper showing what may well be the future of creating video game worlds: an AI that is able to construct most of its own 3D landscapes.
Similar to Nvidia’s work that is able to conjure its own celebrity mugshots, the tech would require only minimal input from a human, who would just have to contribute some basic requirements, draw some lines then let the AI do all the hard work: namely, filling in all the gaps with elevation, ridges and natural-looking rock formations.
As the paper states, this kind of tech would only be the beginning; future research could lead to landscapes being generated entirely by the AI, and for the model to be able to handle more complex environmental features like sand and vegetation.
Comments
2 responses to “AI Is Learning How To Build Its Own Video Game Worlds”
For a long time now I have thought that machines could make infinitely more interesting “grounded” open world maps. Sure you lose interesting hand made designs, but if programmed correctly an AI can design and then populate a world following rules that would allow for “civillised infrastructure” to occur much more naturally.
When a person designs a map, there are set pieces that over ride any sort of logic to the terrain or ecosystem. A machine gets around this by sacrificing pride for the sake of intelligent design.
I really look forward to a time when open world game worlds make sense intead of just being filled with fluff.
Wow, a new low bar for calling something an “AI”. A proper AI would tell you to piss off if you asked it to make you a map.