I sympathize with your idea. The thing is, moving gifted kids to automated learning systems would just trade one serious deficiency for another.
Automated systems require that students be able to self-teach. If a student doesn't understand something, he can't ask a question of the video. So he has to find his answer himself, which would require significant motivation PLUS the skills required to find the answer. It's easy to say, "No, he can ask the teacher," but the teacher will be busy with other kids, because they all need help. My kids attend a school using online learning for math, and the teacher is usually busy with someone else.
I have a PhD and am adept at self-teaching (in part because a major goal of a PhD is to teach a student how to figure stuff out systematically). A lesson I've learned is that 1) success in self-teaching takes a long time and multiple focused information sources, and 2) that the skills required to succeed aren't trivial. By focused information source, I mean 2-3 other textbooks on the subject minimum (including a solutions manual and a teacher's edition with explanatory notes in it). Each source emphasizes different ideas, which is what you need when you don't have a teacher who can explain things. While the web has lots of information, it isn't focused, and you can easily end up going down a rabbit hole and getting lost.
IMO, it isn't reasonable to expect an elementary through middle school aged kid to be even marginally capable of succeeding at this process. Even high school kids would have a) trouble and b) not enough time.
As an example, Khan Academy videos can teach an individual skill, but they don't (and can't) provide the kind of subtleties and connections that a knowledgeable teacher provides. So while I agree that a motivated kid with a high IQ will get through Khan stuff quickly, IMO, the learning will be superficial and algorithm based. Also, in many areas, there's no substitute for a discussion in class.
I suspect that many/most educators are unaware of the challenges I described in the first part of this message, and that the enthusiasm over computer-based learning in coming in part from the Dunning-Kruger effect.
The US has a tradition of looking for new solutions to problems, which is good. The thing is, though, sometimes the old-fashioned way really is better (especially when we've created the problem and don't want to admit it, which is the case with ignoring the needs of gifted kids). We pour resources into special needs kids. This proves that it can be done. So why can't we gifted kids too? (Answer: lack of will). Computers won't solve that problem.
