The work aims to analyze - within the current topics of great interest in artificial intelligence - the problem of defining norms and rules that can be used by the intelligent agents in making decisions. A problem that has roots dating back, so much so that the so-called three laws of robotics were developed by Asimov in 1950, general principles that express basic ethical options, but appear completely insufficient to regulate the specificity of the different domains. The need to give substance to specific systems of rules arises from the pace of technological progress, which continually gives rise to new situations worthy of attention, which cannot always be framed within an already existing system of rules. There is a particularly strong need, then, to give space to an anthropocentric vision which is not limited to allowing the a posteriori discernment of the operations of computer systems in general, and of automatons in particular, predicting that regulatory compliance also plays prior role. This would allow, already in the conception, design and implementation phases of the IT systems, the identification of specifications that take into account the ethical and regulatory constraints applicable to the activities that they will/can carry out, ensuring that the engineering of the software is aimed not only at the needs of application effectiveness and efficiency, but also at guaranteeing its compliance with the aforementioned ethical and regulatory constraints in an ex ante solution for the harmful events.