The concept of "alignment" in artificial intelligence (AI) refers to making sure that AI systems act in ways that match human intentions and values. This means not only creating AI that achieve the goals people want but also ensuring they do not use harmful or dangerous methods to get there. The term has gained attention through researchers like Stuart Russell, who in a 2014 white paper argued that autonomous systems should align their values with those of humans to prevent unintended risks. The idea of alignment can be traced back to the 1960s, when Norbert Wiener, a pioneer in the field of cybernetics, warned about the importance of ensuring that the goals programmed into machines are the ones humans truly want. In French, the term "alignment" is sometimes translated in a way that suggests simply making machines follow instructions, but the concept is broader—it involves creating AI systems that are in harmony with human values and ethical standards.