The Ethics of Artificial Intelligence : an Overview

Since the end of November 2022, the general public has been fully aware of the potential of generative artificial intelligence (AI) tools like ChatGPT. While specialists in the field have been working for several years on language models capable of generating content in a natural way (text, image, sound), the public launch of ChatGPT will certainly go down in AI history as a moment of democratization for tools that, according to their critics, are still too little mastered. At the same time as AI tools are gaining ground, ethical issues are also undergoing a phase of democratization.

These generative AI tools are part of a whole range of tools, methods and techniques grouped together under the umbrella of "artificial intelligence". As defined below, ethics raises the question of values and principles in the design, production, use and regulation of digital tools. These questions of values and principles are being asked everywhere, and the ethics of AI are expanding rapidly. Academic contributions are pouring in at an impressive rate, public debate is seizing on ethical issues and users are taking a more critical look at the tools on offer. Far from being outdone, companies are increasingly integrating an ethical dimension into their internal processes. This growing importance sometimes reflects the deep-seated conviction of management teams, but also an ambition to earn a good reputation by being at the forefront of a movement that we sense to be profound or, at the very least, to avoid accidents along the way that could damage the relationship of trust with customers and partners. Ultimately, companies' initiatives are aimed at anticipating future regulations on the uses of AI. Following the example of the European Union (EU), the regulation under discussion on AI – aimed at establishing harmonized rules for EU countries – will durably alter the way companies, public administrations and individuals design and use AI tools.

This article aims to provide an overview of AI ethical issues for engineering specialists. In addition to a general explanation of the challenges involved, each section concludes with a series of questions directly relevant to companies/organizations in their management of AI tools.

The text is organized into five parts. The first deals with the definitions of artificial intelligence and ethics. The second presents the approach of the navigation tool chosen to structure the questions to be asked. This tool acts as a compass for finding one's way through the ethical questions; it proposes three families of ethical questions, which are developed separately. The third part deals with the first family of ethical issues, namely ethics in technology. Ethical choices in the design and implementation of AI tools are at the heart...