The Ethics of Artificial Intelligence: Challenges, Dimensions, and Prospects for Responsible Use

Abstract

The penetration of artificial intelligence in sensitive sectors such as health, finance, and public administration is advancing apace, giving rise to important ethical and governance problems. Due to fairness, accountability, transparency, and privacy issues (FAT/ML = Fairness Accountability-Transparency/Machine Learning), they are central in the responsible deployment of AI. In reaction, rules and laws like the European Union General Data Protection Regulation (GDPR) have been implemented to safeguard people's rights and to provide guidelines for the ethical use of AI. But the enactment of such laws also presents new challenges related to the availability of data, the accuracy of models, and the performance of systems. In this paper, we conduct a thorough characterization of AI ethics and governance, specifically in terms of fairness, accountability, and privacy, and discuss the effects, such as GDPR, on the performance and robustness of AI models. The results call attention to the costs and benefits of ethical adherence vs. technical performance, and to novel governance paradigms that strive to reconcile innovation and responsible AI use.