Why: AI is going to be the way of the future, and is pretty much one of the most developing and revolutionizing aspects of robotics. Thus, computer science, or similar majors (I heard there will be an artificial intelligence major) should be a major.
Robotics is a multidisciplinary application of science and technology. You need mechanical guys to build it, electrical guys to wire and control it, and computer scientists to do the programming. So any of those majors (and several others) can be “optimized” for robotics. There is even a small but growing number of schools who offer robotics as a major.
Where I work, it is mostly computer scientists who do robotics on a daily basis. We tend to order our robots from a catalog and do the interesting stuff with custom software - and this is a trend that I only see becoming more common going forward. So my recommendation is whatever course of study you pursue, you should at least take introductory computer science courses. As you said, AI is a very young field where we still have much to learn and there is going to be a lot of work in that area in future years.
Like [Jared341] said, robotics is indeed multidisciplinary. You can get into robotics focusing on any of the fields you mentioned (and many more in addition to those: neural science, cognitive psychology, mathematical sciences, etc.), but if CS is your passion, definitely focus on that. But yes, electronic and/or mechanical familiarity will indeed help you when you actually have to work with the mechanical beasts themselves.
Here’s my generic recommendation:
Some of the best advice I’ve ever heard is to not think so much in terms of majors or minors (yes, your college will make you declare one at some point), but instead look at the course offerings and research opportunities that are available at your university and get involved in the ones that really capture your interest. The extra pieces of paper look nice on resumes, but they just say that you took all the courses that somebody decided were necessary to obtain an adequate knowledge of some subject. The cliche is that a lot the knowledge you’ll end up using in your job probably hasn’t even been discovered yet; maybe you’ll even start your own new academic field.
This does seem to be the common thought among programmers and the such. However, I am not convinced yet. Perhaps we will have pseudo AI soon. But I think it’ll be centuries before there is ever anything close to true AI. Partly because of the difficulty, and partly due to politics.
But more on topic: What form of robotics do you want to work in? Animatronics and unmanned vehicles certainly fall into the broad category, but they require completely different focuses.
If you want to work with animatronics, the focus will be on mechanical. How do you make this look and feel real?
If you want to work on unmanned vehicles, the focus will be on programming. We have the car, now how do we make it think?
Robotics is a very broad brush. You might try refining your strokes if you want a masterpiece.