I suppose the benefit of a human system is that we can correct errors that we haven't fully anticipated, by using a general rule such as legal precedent. Remote

I suppose the benefit of a human system is that we can correct errors that we haven't fully anticipated, by using a general rule such as legal precedent. Remote edge cases in a purely machine system would lead to uncorrectable errors, from time to time. I think an ideal system would combine machine rules to avoid dumb-ass mistakes like mislabeling a share cert on transfer, with human systems to overrule unexpected bad results (e.g., transfers that were obviously not intended by the parties, though processed mechanically).