A new article in Scientific American explores how teaching artificially intelligent computers artistic practices may help advance their creative problem-solving. Thinking creatively, after all, will be necessary when AI is introduced more significantly into the real world, as when self-driving cars face unforeseeable problems on the road. The best model for this approach comes from German AI artist Mario Klingemann, who has developed generative adversarial networks (GANs). In GANs, a first network generates images for a second network to analyze, and the former network learns by processing the latter network’s evaluation. This new approach could have major implications for how AI computers make decisions.
Thanks for reading InsideHook. Sign up for our daily newsletter and be in the know.
