The only thing you can trust them to do is screw it up
I expect we're going to see a cultural transformation over the coming decades, which may see the word 'science' itself falling out of use. The sclerotic priesthood of the academies is indeed bleeding prestige in real time, in large part due to its observable inadequacies.
Simultaneously a far more nimble, open-source epistemological network is emerging on the weird corners of the Internet. Publication in peer-reviewed journals is irrelevant to this network. Instead, analysis is pushed out immediately, and then immediately opened to testing and critique by the entire network. "But does it work?" becomes the key question ... really a return to the experiential empiricism that has defined the "scientific method" since the first Australopithecine smashed marrow out of an antelope femur with a chunk of flint.
Rather than holding tenure at large institutions, this network will be dominated by independent scholars supporting themselves via crowd-funding and other market activities. Success will be found by cultivating the respect and interest of a public following, rather than impressing a board of senior academics.
I suspect 'science' as a term might fall out of favor for two reasons. First, it has become weighted down with ideological baggage that is very distasteful to those who seek liberation from the Cathedral; having to constantly distinguish between science and THE SCIENCE™! is an unwieldy terminology. Second, linguistically, the root of 'science' is 'to separate' - to break knowledge down into ever-more granular and isolated categories. By contrast, the imperative now seems to point towards an era of holistic unificationism. That implies an epistemological framework in which the decompositional process of analysis is just one leg of a cycle that must also include synthesis of it is to have any meaning or utility, and science as currently conceptualized is terrible at synthesis.