Main Article Content
Artificial intelligence has in particular changed certain processes. Decisions, which impact both the organization and individuals, can be made by algorithms, including for example granting credit, recommending medical treatment, accessing a service. The implementation of AI-involves new complexity and therefore generates new needs for control and mastery. There is a wide variety of algorithms used for machine learning. They can be grouped into several distinct categories. Algorithms offer many possibilities, but remain simple sets of instructions. They are created by humans, and therefore can have flaws. According to the recommendations of the CNIL1, the principle of loyalty must be applied to all algorithms. Any algorithm, whether or not it processes personal data, must be loyal to its users, not only as consumers, but also as citizens. Rhetorical questions: can we speak of „ethics of algorithms“ and how to apprehend these humanized robots?