Progressive disclosure

Abstract

As we increasingly delegate important decisions to intelligent systems, it is essential that users understand how algorithmic decisions are made. Prior work has often taken a technocentric approach to transparency. In contrast, we explore empirical user-centric methods to better understand user reactions to transparent systems. We assess user reactions to transparency in two studies. In Study 1, users anticipated that a more transparent system would perform better, but retracted this evaluation after experience with the system. Qualitative data suggest this arose because transparency is distracting and undermines simple heuristics users form about system operation. Study 2 explored these effects in depth, suggesting that users may benefit from initially simplified feedback that hides potential system errors and assists users in building working heuristics about system operation. We use these findings to motivate new progressive disclosure principles for transparency in intelligent systems.

Evaluating saliency map explanations for convolutional neural networks

“…Moreover, half of them were shown detailed information about the classi cation scores produced by the CNN. Our results indicate that when saliency maps were available, participants answered correctly more frequently than when they were absent (60.7% vs. 55.1%, p = 0.045). However, the overall performance was generally low even with the presence of saliency maps.…”

“…ey found that this helped participants to decide faster on whether to trust the prediction. Springer and Whi aker [55] conducted two user studies. ey used a system that predicted the emotional valence of participant's wri en experiences.…”

Data-Driven Personas for Enhanced User Understanding: Combining Empathy with Rationality for Better Insights to Analytics

“…Since personas were first introduced, a range of other online analytics platforms, services, and tools have emerged (Cooper, 2004;Springer & Whittaker, 2019) -including Adobe Analytics, IBM Analytics, Facebook Insights, Google Analytics-as well as integrating services (e.g., HootSuite, Tableau, etc.) that organizations can use to understand their users and user segments.…”

Improving Support for Contingency Planning in Air Traffic Flow Management

“…Transparency can increase operator trust and use of the system Springer & Whittaker, 2019). Providing the logic behind system recommendations can increase trust more than providing probabilistic assessments of risk or no explanation at all .…”

“…Providing the logic behind system recommendations can increase trust more than providing probabilistic assessments of risk or no explanation at all . However, the overall body of research is not clear in how best to design transparent systems (Springer & Whittaker, 2019).…”