With respect to specific modelling efforts policy makers sometimes act as antagonists because, without joint modelling of all the concerns they are trying to ba

With respect to specific modelling efforts policy makers sometimes act as antagonists because, without joint modelling of all the concerns they are trying to balance, they are forced to rely on their intuition. This leaves society particularly exposed to the character traits of our leaders, because it is they who end up modelling the interaction effects (in their heads). For example they might claim in the early stages of an epidemic that we actually just have linear case growth. It is then up to the modeller to try to justify why that isn't the case, which in limited-information regimes can be near impossible in any rigorous sense, for the reasons you describe. However, assumptions are important, and showing decision makers what might happen given certain assumptions (and providing a justification for those assumptions), which is part of what a model does, can cause them to reassess their current thinking, and it is for this reason that these kinds of models can remain useful, even if it is not possible to reasonably distinguish between some set of models given available data. Returning to the epidemic example, justifying why exponential rather than linear growth is likely in the early stages of spread is something that can be done on an intuitive level, and explaining this to policy makers, and showing them the range of possible outcomes that are consistent with current (at the time limited) knowledge, can be extremely powerful.