The Role of Algorithmic Audits and Other Soft Law Approaches in Informing Users' Calibrated Trust in Artificial Intelligence Tools

The Role of Algorithmic Audits and Other Soft Law Approaches in Informing Users' Calibrated Trust in Artificial Intelligence Tools

연구 분야: Artificial Intelligence

학회: CSCW Companion '24: Companion Publication of the 2024 Conference on Computer-Supported Cooperative Work and Social Computing

Artificial Intelligence (AI) is becoming more and more a part of our daily life and is affecting a wide variety of users. While there are many beneficial aspects of AI, there are also growing concerns about harms to the public. In addition, there is wide-spread distrust in AI in general in the public. Algorithmic audits could help to build calibrated trust by users. Algorithmic audits are slowly becoming more common, in particular in the Human Resource (HR) sector. Previous research has focused on the use and design of audits but less on how they are received by users' and how they could help users to calibrate their trust in AI-based systems to the trustworthiness of the system. The research described below aims to fill this gap. Based on initial work on algorithmic audits in the HR sector, the research will in the future expand to go beyond this sector and also cover other AI governance mechanisms, in particular soft law approaches such as the NIST AI Risk Management Framework.