General conclusion

At the close of this journey, a few recurring threads stand out. AI is a general-purpose technology, comparable in scope to electricity or the printing press: it cannot be confined to any single sector, and it transforms them all. Its recent history is one of a turning point (the Transformer, large models, agents, embodied AI), driven by a colossal infrastructure (chips, data centers, energy) and shot through with structural tensions that we have encountered from chapter to chapter: capability versus control, openness versus security, concentration versus democratization, and the rivalry between great powers, from the United States and China at the forefront to Europe and the emerging hubs (India, the Gulf, South Korea, Japan).

Faced with a technology as powerful as it is uncertain, neither starry-eyed enthusiasm nor fearful rejection is a good guide. The most useful stance, the one this course has tried to embody, is informed lucidity: understanding the mechanisms, telling fact from hype, weighing promises against risks, and keeping in mind that the future of AI is not written. It will depend, in large part, on the technical, economic, and political choices we make collectively. Understanding AI, as this course sets out to do, is the first condition for taking part in it.

Thus ends the main body of this course. The appendices that follow (glossary, chronological landmarks, and sources) are its reference companions.