Some may remember a similarly controversial topic from 2006, based on a Chronicle article on how Texas Tech's writing program handled online grading.
Online grading like TTU's - and distance education classes, for that matter - have prepared people to be more receptive to the idea of outsourcing or offshoring grading. (To a lesser extent, so has the practice of accepting first-year composition credits from community colleges.) Slowly, education follows the lead of other industries, segmenting specializations and seeking to maximize self-programmable labor while automating or outsourcing generic labor.
Higher education has been insulated to some degree from this trend. But radical tuition increases, budget crunches, and the looming crisis in underfunded pensions will tend to lead stakeholders to consider less expensive solutions. Nobody will think these solutions are perfect, but I expect a lot more "satisficing" than we're used to seeing. Especially in expensive-to-run classes such as those with small class sizes.
The obvious counterargument, of course, is that grading in writing classes is not generic labor: it's problem-solving, it's grounded in particular contexts, it's sensitive to particular students' needs and developments. Sure, of course, I agree. But as a field, we're going to have to start articulating clearly, for outside audiences, what the vital contribution of instructor grading is. Not the preferred contribution, not the best-practices contribution, but the contribution that can't be made otherwise and that is critical to the university's mission. We need to argue that grading can't be made generic.