Perhaps you've seen the announcement from Harvard (also reported in the NY Times) that the Arts and Sciences faculty have voted to create an open access, online repository of their scholarly articles. This led to some discussion on the TechRhet listserv about whether or not the articles were going to be peer-reviewed and accepted by journals before being added to the repository. The reports I've read are not specific, but it looks to me like that will be the case. It will be interesting to see how scholarly journals respond to this shift. Maybe Harvard scholars think their work is so valuable that journals will want to publish it even if it is available somewhere else for free. On the other hand, if Harvard can make it happen, maybe the rest of us can as well. After all, SUNY is a state institution, working for the public; there's a logic in making our work publicly available.
I do think there would be some changes that would need to take place. Most of the significant work of scholarly publication is paid for through faculty salaries. I.e., it's part of our job to do research and write articles; serving as an editor or a reviewer is also an important kind of academic service/scholarly work. Scholarly journals, to my knowledge, never pay anyone for this work. Scholarly journals may pay people to do some line editing or copy editing. They might pay to get layout done. They obviously pay for the printing, binding, and mailing of issues. They handle subscriptions, accounts, advertising, marketing, and various service requests.
The question of copy-editing is a valid one. It's quite possible that in other fields, faculty already pay for some copy-editing assistance. We have student interns who do some copy editing for faculty. Also you would have to have someone who would design and maintain this online repository. You will probably also need to have someone do marketing and handle service issues/tech support related to the website. However, universities already have individuals and departments that do this kind of work.