Irreversibility is the metric for weighing impacts- any other impact has a chance of the world returning

Most worst-case scenarios appear to have an element of irreversibility. Once a species is lost, it is lost forever. The special concern for endangered species stems from the permanence of their loss (outside of Jurassic Park). One of the most serious fears associated with genetically modified organisms is that they might lead to irreversible ecological harm. Because some greenhouse gases stay in the atmosphere for centuries, the problem of climate change may be irreversible, at least for all practical purposes. Transgenic crops can impose irreversible losses too, because they can make pests more resistant to pesticides. If we invest significant wealth in one source of energy and neglect others, we may be effectively stuck forever, or at least for a long time. One objection to capital punishment is that errors cannot be reversed. In ordinary life, our judgments about worst-case scenarios have everything to do with irreversibility. Of course an action may be hard but not impossible to undo, and so there may be a continuum of cases, with different degrees of difficulty in reversing. A marriage can be reversed, but divorce is rarely easy; having a child is very close to irreversible; moving from New York to Paris is reversible, but moving back may be difficult. People often take steps to avoid courses of action that are burdensome rather than literally impossible to reverse. In this light, we might identify an Irreversible Harm Precautionary Principle, applicable to a subset of risks.' As a rough first approximation, the principle says this: Special steps should be taken to avoid irreversible harms, through precautions that go well beyond those that would be taken if irreversibility were not a problem. The general attitude here is "act, then learn," as opposed to the tempting alternative of "wait and learn." In the case of climate change, some people believe that research should be our first line of defense. In their view, we should refuse to commit substantial resources to the problem until evidence of serious harm is unmistakably clear.' But even assuming that the evidence is not so clear, research without action allows greenhouse gas emissions to continue, which might produce risks that are irreversible, or at best difficult and expensive to reverse. For this reason, the best course of action might well be to take precautions now as a way of preserving flexibility for future generations. In the environmental context in general, this principle suggests that regulators should proceed with far more aggressive measures than would otherwise seem justified.