This kind of data archiving would help reproducing data analysis. It's often the case that data collection is hard to reproduce. I suspect that, at least for some articles, it would be worthwhile to attach video of the experimental methods being performed to an article. That way, everything involved in generating the raw data will be documented.

I don't think this will ever be commonly done. And even if investigators were willing to do it now, the bandwidth and storage requirements would probably be too great. But at some time in the future I expect that it will be done, especially in the event of a dispute.