Examples of problematic AI applications

Abstract

This article presents three examples of problematic applications of artificial intelligence (AI). These examples, "inferring personal characteristics", "autonomous weapon systems" and "computer-assisted early warning and decision-making systems", are about potentially significant impacts for humanity. However, the author is convinced that most AI applications are positive for humans. In order not only to take a negative view of AI developments, but also to provide indications of recognition criteria that can be helpful in assessing problematic applications, three applications with their own involvement are also presented. The aspects of vagueness and uncertainty, which are also relevant for the problematic AI applications, are also addressed.