DESIGN OF DIGITAL APPLICATIONS ORIENTED BY THE USE OF LEARNING ON HUMAN-CENTERED MACHINES (HCML): CONSIDERATIONS ABOUT USER EXPERIENCE IN GAME STREAMING PLATFORMS FROM THE TWITCH REDESIGN PROCESS

Abstract

The use of Machine Learning (ML) has been increasingly present in several applications, especially in content recommending systems available on streaming platforms. However, sometimes the automation provided by machine learning can negatively impact the user experience, making the user no longer in control of which content he wants to access, generated from a vicious cycle of recommendations based on preferences of use. Human-Centered Machine Learning (HCML) aims to provide alternatives so that machine learning (ML) can allow explicit control to the user, as opposed to an implicit interaction. This article aims to demonstrate the potential of the applicability of HCML and optimization of navigation in the redesign of Twitch, a platform for streaming electronic games. Thus, an analysis of the platform interface was carried out based on the navigation design, identifying possible alternatives adapted from the premise established by HCML.