There are far, far, far more instances of people failing to respect the power of science as the far-and-away best way to know something, if you can apply it decently, than of people thinking it's the only way to know something.

The point about human biases is true (not sure Kuhn is the right citation; he's more about incommensurate belief systems), but the reason it's a problem is because that disrupts the operation of the scientific method. Indeed, much of the problem here is, ironically, in people failing to respect the power of science as the best way, and falling back on other ways to know something!

People say "science!" but instead rely on authority (of a person who happens to be a scientist). People say "science!" but instead act like difficult matters with ambiguous evidence are proven to be one way and not another. People say "science!" but instead rely on tradition that they never seriously tested (there is just "no evidence" that it is wrong because they didn't bother). People say "science!" but instead mean that you have to rise to the level of subjecting your ideas to a blizzard of falsification attempts with much evidence in support, while they can cling to their prior beliefs with at best poor evidence.

Actually internalizing the lessons about epistemology that science teaches us by example--by better grasping that elusive but ever-so-potent way to know things--is still by far our best way to avoid error, and still has enormous room for improvement.