Yeah, in my heads, human creation or curation is a "necessary, but not sufficient" thing. A ton of junk is already out there pre-computer-generated-content. So

Yeah, in my heads, human creation or curation is a "necessary, but not sufficient" thing. A ton of junk is already out there pre-computer-generated-content. So being human created isn't enough to let us know it's high quality - but not being human created is enough for the immediate future to let us know that it can't be trusted for domains where you aren't expert enough to tell the difference between "plausible" and "correct." And then in the anonymous public internet, you simply won't be able to tell "human created" from "computer created" so you'll have to just treat it all as low-value spam. Another interesting question here: let's say someone created a future AI that could actually be relied on to give "correct" or at least "properly caveated and nuanced" answers to questions - how is the layman possibly to know if that engine is actually being used behind the scenes of any online text? E.g. does the bullshit ML-generated content poison the well for "better" AI too, not just for anonymous people?