Human bias is a whole huge area apart from just quality, not sure we want to get into that. I mean is it human bias, or is it reality bias. e.g. Poland hasn't had any terrorist attacks (or very few?), is that human bias, or is it reality?
I definitely don't treat its output with full trust, but I've been pleasantly surprised that even when I give it bad or incorrect guidance (unintentionally) it has caught my mistake, corrected me, and I've learned things.
For the QA case, I suppose what you're getting at is that if it can't be fully trusted, you might get incorrect QA results -- false negatives, false positives -- I'd agree but I think you just have to find an effective way to use the tool. Perhaps the obvious way most people would want to use the tool, is not in fact the best way to use the tool.
But just because the tool isn't delivering the perfection people hope for, doesn't mean there isn't some other way to use it that still catches (some) mistakes and adds value.
I definitely don't treat its output with full trust, but I've been pleasantly surprised that even when I give it bad or incorrect guidance (unintentionally) it has caught my mistake, corrected me, and I've learned things.
For the QA case, I suppose what you're getting at is that if it can't be fully trusted, you might get incorrect QA results -- false negatives, false positives -- I'd agree but I think you just have to find an effective way to use the tool. Perhaps the obvious way most people would want to use the tool, is not in fact the best way to use the tool.
But just because the tool isn't delivering the perfection people hope for, doesn't mean there isn't some other way to use it that still catches (some) mistakes and adds value.