Individually these units could be pretty solid, but together I think they may be a little much for a one-year course.
I'm a little unclear about the relationship between 2d and 3d through the course. But nonetheless, my own advice is to stick to 2d throughout. If you're worried they'll think that's uncool, consider framing the course around mobile game development instead; I think it'll keep expectations more realistic. They'll be used to 2d mobile games with achievable pixel or vector art, and with a relatively simple core interaction loop. (NB: I think GameMaker requires an extra license fee if you want to actually compile to a mobile platform. But saying "We're doing simple mobile-style games" will at least give them an idea of the scope of possible projects. When I made my first game at university, the prof said "We're doing 1980s-style arcade games"; we didn't actually target an actual vintage computer, but it kept our expectations realistic.)
Unit 1: I'm all for teaching game design by getting out the pens & scissors. Just a note that implementing *novel* game designs as software does require somebody to do the programming at some point. You can't really make novel game designs (in the technical sense of design) by checking boxes and copy/pasting code from tutorials; you can only do minor variations of standard designs. Just something to keep in mind for point 4. If someone creates something amazing in Unit 1 and wants to bring that forward into Unit 4, what means do you have to support them?
Unit 2: Creating a 3d model is fun and perhaps doable, but it's a big leap from making a model to using it in a game in Unit 4, unless you want it to bounce around rigidly like a scarecrow. I'd suggest instead creating a 2d character and teaching them some of the timeless basics of traditional animation (like the 12 Principles of Animation from Thomas and Johnston).
Unit 3: I think a good way to draw in art history is to point out the parallels between the modernist period and today's videogame landscape. Videogame art is in some ways in every period at once, but I'd say we're more in our modernist period than anything else. The state-of-the-art has gone so far towards photorealism that a lot of devs are stepping back from that arms race and asking "Ok, what *else* can we do? What are the limits of this medium in other directions?" This is partly practical, but it's also an exciting time of artistic exploration, and debate, and feuds, and manifestos. What luck, that your students are around to see, and participate in, a modernist period! If they realize that, they may appreciate the original modernist period more ;D
So give them a balanced diet of art games, especially to get their mind out of the place where "photorealism" is their default artistic choice.
This might be something to draw through the whole course, and then have the third unit be more about environmental, level, and narrative design.
Unit 4... It depends in part on what you mean by "silly" and what you mean by "violent". Much of what an inexperienced non-programmer can make in a person-month is silly. And your interpretation of "nonviolence" could mean either "no machine-gunning people with blood flying everywhere" (a reasonable request for a school project!) or "compatible with ahims?" (which is a lot to ask of someone designing a game by putting together code fragments from tutorials).
And I should note that I only design nonviolent games, so I'm more likely than most to be sympathetic to that final project. Retrofitting existing game designs to be nonviolent is my idea of great fun. But I don't think it's a good fit for the students' likely skill-set, and unless you're already a game designer and programmer you won't be in a good position to help them.
In all, you might want to consider something that's not so much a game design course -- any true design they might do requires some programming to implement-- and instead concentrate on the game development part (character, animation, and environmental design). Maybe if you already had a game design and implementation that fit your existing creative demands, and had each student or team design a character and level.
For example, say the theme of the game is "A teenager overcoming a real-life obstacle that manifests as a videogame challenge", and you have a run & jump & avoid obstacles game already implemented (a completely standard set of mechanics that you could piece together from tutorials). The student designs an animated avatar and a level that, through its environmental and narrative design, serves as a metaphor for the real-life obstacle, and MAYBE designs a novel mechanic if they can program or are willing to learn. That gets them concentrating on design in the artistic sense ("How do I best represent my theme in the chosen medium?"), which is probably more your forte, and it's not "silly" "violent" stuff, and it also doesn't so much run into the problem of students designing games more ambitious than they can actually pull off.