Transparancy concerns affect how people perceive and trust AI

Explain how the algorithm works so I can trust it!

March 16, 2023

New research states that transparency concerns affect the fairness, accountability, and privacy perceptions of algorithmic decision-making systems (ADM) relating to trust and use.  

The Dutch Government's childcare benefits scandal painfully illustrates the risks associated with automated decision-making. Yet the use of algorithms has become indispensable. Companies, governments, and the nonprofit sector increasingly use these systems to make automated decisions impacting our daily lives. The public and policy-makers are continuously discussing concerns about the transparency of these systems. It is unclear if the public is really concerned and if it affects their adoption of algorithms.  

New research by Banu Aysolmaz makes it clear that users base their perceptions of an algorithm's fairness, accountability, and privacy on their transparency concerns, impacting their thoughts about the trust and usefulness of these systems. The model was tested by a large survey among 2500 Dutch citizens considering five algorithmic uses. The research was published in Telematics and Informatics Journal

Banu: “The insights from this study can help organizations improve the adoption of their ADM systems by addressing relevant transparency concerns. We need more research on the ethical concerns and design of relevant transparency mechanisms that can facilitate the specific needs of diverse users. ”