The effects of user-AI co-creation on UI design tasks

Abstract

With the boost of GPU computation power and the developments of neural networks in the recent decade, a lot of AI technique are invented and show bright potential of improving human tasks. GAN (generative adversarial network) as one of recent AI technique has powerful ability to perform image generation tasks. Besides, many researchers are working on exploring the potentials and understand user-AI collaboration by developing prototype with the help of neural networks (such as GAN). Unlike previous works focus on simple sketch task, this work studied the user experience with UI design task to understand how AI could improve or harm the user experience within practical and complex design tasks. The findings are as follows: multiple-hint AI turned out to be more user-friendly, and it is im-portant to study and understand how AI’s presentation should be designed for user-AI col-laboration. Based on these findings and previous works, this research discussed about what factors should be taken into consideration when designing user-AI collaboration tool.