Assessing the trustworthiness of artificial intelligence

Roberto V Zicari

INTRODUCTION

This chapter illustrates how to evaluate whether an artificial intelligence (AI) system, used, for example, as a decision-making tool, is trustworthy. Using machines to make decisions raises several ethical considerations for society. AI ethics has increasingly focused on converting abstract principles into practical action. This chapter describes a new process for assessing trustworthy AI in practice, called “Z-inspection”.¹¹Z-inspection® is a registered trademark. Z-inspection® is open access and distributed under the terms and conditions of the Creative Commons (Attribution-NonCommercial-ShareAlike CC BY-NC-SA) license. For reasons of textual clarity, it is not the publisher’s house style to use symbols such as ®, etc. However, the absence of such symbols should not be taken to indicate absence of trademark protection; anyone wishing to use product names in the public domain should first clear such use with the product owner.

AI,²²Artificial intelligence refers to systems that display intelligent behaviour by analysing their environment and taking actions (with some degree of autonomy) to achieve specific goals. AI-based systems can be purely software-based