I would flip this around and say that we must establish patterns of behavior that make trustworthiness readily apparent as widely as possible. Otherwise our trust is blind, and will probably be abused sooner or later.

At the root of this trust is a correspondence between belief and reality, but the "trust but verify" step isn't considered as important as it should be, and the "yes, reality, please" consensus is far too weak.

For instance, I think a good fraction of the Covid mess is the result not of cult-minded antivaxers, but a completely counterproductive reaction to cult-minded antivaxers. The reaction was to clamp down on things and belittle-and-shame rather than expansively share what we know and don't and why. The clamping lost a lot of people who are equipped to know better, and could have dragged the others back.

One of the most egregious examples was when attention started shifting from HCQ to ivermectin. Well before there were sizable studies that were conclusive, the supposedly science-following people were belittling ivermectin on the basis of, as far as I could tell, it being picked up by the HCQ people (plus very shoddy evidence that shouldn't have convinced anyone either way--I read the early papers); and after better results came in that at least documented clearly that ivermectin wasn't a miracle, the anti-ivermectin people then doubled down by mocking people for taking "horse dewormer", despite ivermectin being a critically important drug for fighting human nematode parasites, too.

Anyone who was interested in ivermectin early on could be forgiven for interpreting the response as thought-controlling you to funnel profits to big pharma, and practically everything afterwards reinforced the notion.

This is one of oodles of examples. We used to hope that institutional trust, plus institutional safeguards to make the trust mostly worth it most of the time, would be the mechanism that we need. But that is being eroded on all sides (both the actual quality gets eroded, but also the perception gets eroded more), so we need alternate schemes (like doubling down on transparency and openness) to compensate.