Each of those fields is huge, so the best way for you to get an idea of what they involve is do some research on your own. But here are a few short blurbs on each:
- HCI & Graphics: Usually HCI and graphics are fairly disjoint. However, HCI involves learning a lot about your target audience, creating and running studies, etc, and then iterating over possible designs (and iterating some more after that). Graphics involves a lot of math (linear algebra), and otherwise working with existing tools (i.e. game engines, animation software). You don't have to be particularly artistic, although it certainly helps. Creativity is more of a big deal here, although that could also apply to most of computer science.
- AI: This is really a gigantic field, but two of the biggest focuses nowadays are machine learning and natural language processing. Lots of math (linear algebra and others).
- Security: For cryptography, lots of math (number theory and others). Otherwise, it really depends on the domain. Network security requires knowledge of networks. Database security requires knowledge of databases. Systems security (ironic that I use this, as you'll see below) requires knowledge of your system.
- Systems: To be honest I don't even know what this means. Distributed systems? Embedded systems? Mainframes?
As you can see, narrowing this down to something helpful is not easy.
If I were you, I would also consider Software Engineering, although that just a personal bias. It's really important in industry, although not exactly as taught in university. Do you have to specialize? I've always thought it was a bit ridiculous to expect students to choose a specialization coming into the program, before they've even been exposed to many of the involved areas. You'll certainly know a lot better after, say, 2nd year.
