Abstract:

The exponential growth of Artificial Intelligence applications in the last decade, led to a debate in the science, business, political and activist communities about the ethical implications such systems. The dispute about ethics in AI is not new, already 60 years ago in the birth hour of Artificial Intelligence, the discussion on unwanted moral and technical consequences of automation has begun [1]. In the last decade, research on ethical aspects of Artificial Intelligence has increased tremendously. With ethics as a vast domain on one side and Artificial Intelligence with its broad and growing complexity on the other, the factorization of these areas cannot be approached in the entire spectrum. In this article we would like to introduce a conceptual model of ethics assessment for Artificial Intelligence systems based on tabular data.
At first, we present and reevaluate a model of ethical dimensions after which we will analyze contextual differences and risk levels of Artificial Intelligence systems. Based on the context and the risk level we derive an ethical assessment to be applied at every stage of the development lifecycle of an Artificial Intelligent system based on tabular data.