There have been an increasing number of statements and publications relating to developments in — and the prospects for — artificial intelligence (AI) in recent months, dangling before our eyes major changes in our daily lives and in many sectors of activity. As is often the case, this kind of scientific breakthrough may seem to imply either the best or worst of possible futures, but it remains also very greatly dependent on what individuals will find acceptable. As he rehearses here some of the limits to the roll-out of AI, Hubert Landier reminds us that not everyone — in France, Europe or the developing nations — has the same expectations of these developments, since there is a real difference between human and artificial intelligence which technology is not yet able to overcome. And there is an even greater difference between the dreams of the developers in Silicon Valley and the reality of our civilization, as it faces up to the physical limits of the planet and the facts of uneven development. This is a viewpoint which draws on the current debates on AI and tends toward downplaying the expectations around it.