The precautionary principle, that proposed changes should not be implemented unless it has been demonstrated that they will lead to improvement, has become a mantra of modern decision making, ranging from scientific and environmental developments to organisational management. In the context of climate change, Softest Pawn argues that it wrongly applied and flawed in any case. I don't agree in detail, but it has become such a commonplace that it is worth exploring some more conceptual aspects of the way it is used.
It assumes that the situation is stable
If the choice is change or no change, it is reasonable that the case for change should be robust. But in many contexts, this is not the choice being faced - rather it is change A or change B, or change a little or change a lot. The PP is no help here - the competing arguments must be considered on their merits.
It assumes that the current situation is acceptable
If the current solution is not resulting in the desired outcomes, then there is no reason to prefer it to changes which may offer better outcomes.
It assumes that timing is not critical
The PP is basically a holding position - the case for change requires more evidence or study, after which the question can be revisited. If the change is time-critical, the opportunity may have gone.
It assumes that the effects of both choices can be predicted
Sometimes they can't, or not accurately, in which case deciding which is 'safest' becomes problematic.
It places the burden of proof on change
A higher level of evidence may be demanded for change than for stability, illogically.
It arbitrarily favours the current situation
Because greater effort is required to initiate change A, judgement is balanced in favour of the status quo (B) - but if the situations were reversed, then option A would be preferred on the same evidence.
So the next time someone says 'better not do anything, to be on the safe side' you may well be able to argue that this is not the safe side at all.