Comment by SkyBelow

Humans are imperfect, but this comes with some benefits to make up for it.

First, we know they are imperfect. People seem to put more faith into machines, though I do sometimes see people being too trusting of other people.

Second, we have methods for measuring their imperfection. Many people develop ways to tell when someone is answering with false or unjustified confidence, at least in fields they spend significant time in. Talk to a scientist about cutting edge science and you'll get a lot of 'the data shows', 'this indicates', or 'current theories suggest'.

Third, we have methods to handle false information that causes harm. Not always perfect methods, but there are systems of remedies available when experts get things wrong, and these even include some level of judging reasonable errors from unreasonable errors. When a machine gets it wrong, who do we blame?