Search

The international conference “Pour des intelligences artificielles au service du corps vulnérable : Les contreforts de l’éthique et du droit” held in December 2021 at the Université catholique de Lyon, in France, explored whether artificial intelligence (AI) be a lever to restore equality between poor and rich, women and men, the disabled, the elderly? In the context of the Covid-19 pandemic, this question is being asked everywhere, including various international bodies, such as the United Nations (UN) and the International Communications Union (ICU). This question requires an impact analysis (leverage) of AI that can generate more social justice (reduction of inequalities). How can we produce such an analysis so that AI can benefit everyone without excluding citizens from marginalized communities? The difficult ethical problem to solve is the following: how far does the impact of AI maximize justice by reducing inequalities? Here the analysis relies on knowledge of AI features that can have a real or negative impact. The purpose of this paper will be to show, based on our framework for analyzing technology impacts and various published texts, the strengths and weaknesses of AI, to answer this question. Applying this framework, we proceed as follows: 1) situate the complex problem of inequalities: those that are generational and socio-economic in nature to which AI has little impact; inequalities to which AI can have an impact, including digital inclusion and social inequalities; 2) situate the terminology of AI: weak (algorithms) or complete (decisional and autonomous) AI; 3) do an analysis of inequalities: digital inclusion and social inequalities showing the strengths and weaknesses of AI. [ABSTRACT FROM AUTHOR]

In what way are chatbots new resources for humans? What are the conditions for their successful implementation? How do chatbots transform human activity? To answer these questions, this paper investigates how four chatbots were implemented and utilized in a professional context based on real user activity. We first show that the current design approach does not take sufficient account of the real activity and its multiple determinants, or of the users' point of view on their activity. Yet these interactive devices participate in a pre-existing socio-technical system made up of a diversity of subjects engaged in activities with multiple purposes. Through the prism of instrumental geneses, we therefore propose to identify how these chatbots can help or hinder human activity, so as to document, for design purposes, the conditions leading to the emergence of chatbot uses. After a synthetic presentation of the empirical results, we discuss the relevance of the instrumental approach when characterizing the contributions and limitations of chatbots. Finally, we argue that a design perspective supporting the postulate of human-machine asymmetry is prolific in supporting a design approach that is more focused on human power to act than on innovation. [ABSTRACT FROM AUTHOR]