An important aspect of Artificial Intelligence (AI) Systems is their User Experience (UX), which can impact the user's trust in the AI system. However, UX has not yet been in the focus of AI research. In previous research, we have evaluated the UX of the Meta AutoML platform OMA-ML, uncovering weak points and proposing several recommendations for ensuring a positive UX in AI systems. In this paper we show that implementing those recommendations leads to measurable UX improvements. We present the UX-improving features implemented in a new release of OMA-ML and the results from a second UX evaluation. The UX of OMA-ML could successfully be improved in four interactive principles (suitability for the user's tasks, self-descriptiveness, user engagement and learnability). We argue that an iterative approach to UX potentially leads to more human-centered AI.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.