> Most Americans

They asked 174.6 million people?

you should ask your flatterbox about random sampling

probably just a ChatGPT

j/k, but I'm pretty sure you could substitute "AI" with a few other keywords here that a lot of people use/depend on: Govt, Healthcare, Social Security, Airport security, heck maybe even science.

The real question is how do you scale something without eroding trust. Transparency has to be part of it but I doubt that it's the only piece of the puzzle and no matter how good your intentions are, there are always people that will refuse their trust (I'm not judging, it's just a fact). As a distributed systems person, I think systems in general work best when they can deal with mistrust and people choose to rather than being forced to use your system to solve their problems. AI is not there yet.