The problem with humans reviewing AI output is that humans are pretty shit at QA. Our brains are literally built to ignore small mistakes. Digging through the output of an AI that’s right 95% of the time is nightmare fuel for human brains. If your task needs more accuracy, it’s probably better to just have the human do it all, rather than try to review it.
I’ve often thought that America suffers from being the first successful iteration of our style of government. It was great and a huge improvement over all the other examples at the time. So much so that much of the world eventually followed in its footsteps.
But where other countries looked at our first successful attempt and further improved and refined the idea, we’re still stuck on that very first version. What was once a radically new idea that worked so much better than everyone else, is now an old, outdated and barely functional relic. We’re the early prototype iPhone 3g, while several other countries have iPhone 6/10/etc