The technical skills of programmer John Carmack helped create the 3D world of Doom, the first-person shooter that took over the world 25 years ago. But it was level designers like John Romero and American McGee that made the game fun to play. Level designers that, today, might find their jobs threatened by the ever-growing capabilities of artificial intelligence.
One of the many reasons Doom became so incredibly popular was that id Software made tools available that let anyone create their own levels for the game, resulting in thousands of free ways to add to its replay value. First-person 3D games and their level design have advanced by leaps and bounds since the original Doom’s release, but the sheer volume of user-created content made it the ideal game for training an AI to create its own levels.
Researchers at the Politecnico di Milano university in Italy created a generative adversarial network for the task, which essentially uses two artificially intelligent algorithms working against each other to optimize the overall results. One algorithm was fed thousands of Doom levels which it analyzed for criteria like overall size, enemy placement, and the number of rooms. It then used what it learned to generate its own original Doom levels.
The second algorithm studied the same man-made Doom levels as the first did, but used what it learned to compare the new AI-generated levels to the originals. The idea is that if you can fool the second algorithm into thinking a computer-designed level was actually made by a real gamer, then it must be as good.
Few gamers are still playing Doom these days, but this approach to AI as level designer is still in its infancy. The algorithms used could also be trained on levels from any first-person video game released over the past 25 years. The more source material it studies, the better the potential results will be.
Procedurally-generated games have existed for a while now, where the software itself continuously spits out unique levels as a gamer plays, but they lack the complexities and nuances that man-made levels deliver. There’s a reason monsters were occasionally hidden right behind doors in Doom, to add to the experience and keep players engaged. But now that an AI can do the same thing, we might one day see games that can generate an infinite number of levels, so they never run out of replay value.