In case of graphic design, school should let the creativity of the children involved unfold. So now you have to prove that GIMP and Inkscape are detrimental to creativity.
The "skills" gained in school do not matter. But open source has another strong poit: convenience.
Compare two stories. A teenager took a course in design on CS2, liked it, but didn't have the software on his home computer so he didn't practice extracirricularly. After coming to the design field in ten years he discovers he doesn't remember anything and the skills he does remember are obsolete. He's also not very good since the lack of practice.
A teenager took a course in design on GIMP. He installed GIMP at home, designed every day religiously for ten years. After coming to the design field he picked up some Photoshop skills and was happily designing ever since.
Anyway, GIMP is seriously limited when compared to the feature set of Photoshop. It makes no sense for him to kickstart a design career using inferior software, just because it happens to be open sourced.
The feature set is an interesting question but an orthogonal to one we've been discussing, that is: preferring the current mainstream, business-adopted software even if it costs significantly more.
I don't see why you demote this to merely an option. It's the cruical selling point. Open source programs also tend to be cross platform (available to more students) and support open data formats (their data will not be locked, flowing freely).
It can be more useful to train students in commercial software if they are ubiquitous everywhere else. For example, in a graphic design course, one would teach students to use Photoshop, Illustrator, InDesign, etc. It would be detrimental to train them in GIMP and Inkscape, as aside from these software being substandard for serious use, hardly anyone uses them. Whereas showing someone how to use even the older CS2 version of Photoshop, these skills will transfer easily to the latest CS6.