Of course, such a system would also have to carry some type of differentiation; for example, if I'm reading at a 3rd grade level, say 2 years behind or ahead of my peers, then it'd be silly to make me read the same books as they're reading if those books either won't challenge me or will challenge me too much.
I guess this would also entail getting rid of class grouping by age, and instead group by ability (and probably at least one or two other salient features). It'll also require that we re-think grades and what they represent, as simply completing a set amount of work doesn't necessarily mean that progress is being made.
The current system in place seems to be geared toward evaluating schools. Since this is proving insufficient, I suspect the next move will be to evaluate classes (Value-added metrics, for example). This one'll prove insufficient, too, though it will take another 5-10 years.
I was thinking today about the potential of digital textbooks/instruction especially in subjects like math. Schools group so many students so arbitrarily that most kids are not engaged at all in their class. The teacher gives a lesson but half of the students didn't understand yesterday's lesson so they don't understand today's either, one third of the students are ready for something new.
Give every student a laptop with extremely limited internet access. Students rotate from class to class getting mostly individual help from teachers on different subjects and work on their individual progress on a web-based teaching curriculum. All Johnny or Suzy has to do each day is make progress towards his/her educational goals, stay on task, and not break anybody or anything and they can be "successful" in the most modest sense. Somebody breaks a laptop, oh well. No work is lost, it's all web based, get a new machine and move on. Smells like the future to me.
The Khan Academy is a prime example of how to use technology well for learning, particularly in instances where there is a "right" answer (such as math, at least how it's taught now).
The danger from this is that the interactivity is lost; I can assign the questions from the end of a chapter in a textbook (which requires no more original work on my part than assigning questions in a digital textbook), but if things are checked by computers then they're likely things that are easily checked.
When I taught, I used those "objective" questions as launchpads, Bloom's Taxonomy-style (one cannot understand "Romeo and Juliet", for example, if one thinks that they lived happily ever after).
A combination between this and some type of small-group/1-on-1 interaction (and even some larger discussion groups) could be extremely powerful.
Given how much curriculum feels like a list of checkboxes to be ticked off (and in many cases, given how it is a list of checkboxes to be ticked off), computerizing the brute, low-level stuff would allow more time for the high-level stuff, particularly if all we're concerned about is rote memorization and recitation.
There are more ways to branch out from there that foster critical thinking and creativity, but this would be a start. This would be a very definite start.