There is a saying 'too many cooks spoil the broth', and it is quite fitting here. As said a small number of experienced / skilled developers can often achieve a far better result far faster than a large number of intermediate level developers, let alone beginners.

The situation is kind of similar to the problems with multithreading. Some jobs are inherently scaleable. If you need potatoes harvesting from a field, you can add workers, and that field will be harvested faster almost linearly as you add the workers. Other jobs are very difficult to scale, particularly programming, as the parts need to work coherently together, and there is a big danger of everyone treading on each other's toes. A large part of software engineering is trying to make modules as self-contained and independent as possible, but this is not a magic bullet and there are limits to scalability.

Artwork is more likely to be scalable imo, however, skill level is very important for artists too, and there is need for art direction and a coherent vision. An artist needs to be trained to the level so that they can produce assets of high enough quality, both technically and artistically. Beginners are unlikely to be able to produce anything usable.

The job potentially easiest to scale imo would be content creation in terms of simple map editing, asset placement and simple scripting. Providing the tools are idiot proof enough. But this again would depend on careful management / direction, and quality control.

As Bregma points out the show stopper in terms of anything commercial may well be intellectual property / legal issues. It may well be easier to invest in procedural map generation tools / helpers than deal with the myriad of issues of employing / directing large numbers of people.