In a cavernous warehouse north of New York City, a 16-foot robot outfitted with a cutting tool etched intricate grooves into a faceless marble head atop an alien-like torso. Water sprayed into the air as an image created with artificial intelligence entered the physical world. In February, during a three-month stint as OpenAI’s first artist in residence, Alexander Reben gained early access to the startup’s Sora text-to-video tool, which instantly generates videos up to a minute in length from written or spoken prompts. Reben, a technologist trained at the Massachusetts Institute of Technology, used Nvidia’s neural radiance field technology to turn Sora’s AI-generated imagery into 3D models.