The GauGAN image creation system Nvidia presented this week is an impressive foreshadowing of AI's rapid advancement into creative fields, like art and architecture—and a frightening example of the increasing ease of producing (nearly) photorealistic inauthentic imagery.

The possibilities of this crude demo tool for future, more sophisticated architectural application are hauntingly plentiful, and the company already hints at exactly that in its announcement: "GauGAN could offer a powerful tool for creating virtual worlds to everyone from architects and urban planners to landscape designers and game developers. With an AI that understands how the real world looks, these professionals could better prototype ideas and make rapid changes to a synthetic scene."

Demonstration of Nvidia's GauGAN system. Video via Nvidia on YouTube.

Since neural networks generate creations based on the set of images they have been trained on, will the creative role of architects soon be diminished to curating the 'styles' of these sets in the hope of somewhat influencing the machine's output?