The public was much more likely to say AI would harm them than benefit them.
There are so many things called "AI" these days, that studies like this are basically meaningless. I think (hope) most people's views can't be reduced to a single binary question.