If something appears to have a potential for harm then we should not wait until we are certain that it will cause harm before doing something about it. For example, if scientists are uncertain as to whether global warming is caused by carbon emissions; this does not mean that carbon emissions should not be reduced.
This principle can be difficult to apply. How certain do you have to be that there is a potential problem? How do you balance the risk of not taking action against the risk of problems the action might create? Will being over cautious stifle innovation?
Do the results of an action justify the means used to obtain them? Sometimes people agree on what is a desirable outcome but disagree about how to get there. For example, would cutting carbon emissions by banning all forms of transport be going too far? A case of using a sledge hammer to crack a nut?
Inaction may cause more suffering than action. For example, it would be wrong not to call out a warning when you see a fast car approach your friend crossing the road. Both actions and inactions have consequences and so may need to be justified.