Mario may be super, but even he must get bored hurdling the same Goombas and falling off the same cliffs over and over. Fortunately, A new artificial intelligence algorithm for what’s called procedural content generation can endlessly produce new levels—and make sure they meet certain criteria. Video games such as No Man’s Sky use procedural content generation to automatically generate up to 18 quintillion unique planets for players exploring the galaxy, a daunting task for any human designer. But programmers still need to hand-craft the rules that tell the computer how to create such content. Game levels are particularly tricky to generate because small changes can make them unplayable – a stray wall can seal off a critical passage —but machine learning, an AI technique by which computers learn from many examples, has generated levels for several games, including Super Mario Bros, Starcraft II, and The Legend of Zelda. In this example of an AI making Mario levels there are two phases. In the first, a “generative adversarial network” learns through trial and error to transform strings of numbers into levels indistinguishable from human-created levels. A second phase then helps find number strings that lead to levels that are not just realistic but that fit certain requirements— such as having a lot of enemies or jumps, giving researchers precise control over difficulty. The researchers believe their approach would work for other games, too. Another method uses adversarial networks to produce new maps for Doom, the classic first-person shooter. The algorithm creates Doom maps that match human-created ones visually and on certain higher-level features, such as the balance of large and small rooms, Procedural content generation doesn’t just save designers time and Mario from tedium. It could also improve video games — allowing games to adapt to players on the fly, so each level is not too hard, not too easy. Just super.