The concept of the precautionary principle came out of environmentalism, but is now phrased more broadly. For example, it is defined by UNESCO as:
"When human activities may lead to morally unacceptable harm this is scientifically plausible but uncertain, actions shall be taken to avoid of diminish that harm."
While this principle is often phrased to include only humans and/or the environment I feel it is equally important in the field of animal welfare. Many of the things we do to animals (such as surgery or close confinement) might reasonably be considered to cause some harm, although the scientific evidence is often missing, contradictory or ambiguous.
If we let lack of data always translate into lack of action this suppresses animal welfare research. The status quo becomes easy to protect simply by not providing the resources or access necessary to carry out definitive research. If, instead, we say that a reasonable expectation of harm obligates us to find a way to end the practice, those wishing to defend the questionable practice are obliged to present data proving harmlessness to continue to use it. The power balance shifts from those wanting to continue business as usual, to those calling for an end to a plausibly harmful practice.
And this obligation goes beyond showing an immediate need for the questionable practice, for example that beak trimming is necessary to prevent the even greater harm of peck injury and cannibalism. The precautionary principle clarifies that the use of a lesser harm to prevent a greater one is not ultimately satisfactory, and that progress must be made towards a system where our method of caring for these animals does not predictably cause significant harm to the animal--because even when the immediate cause is a conspecific, the ultimate cause is a breeding/housing situation that causes the animals to severely injure and kill each other. And removing the beak, while it mitigates the harm in the short term, is not solving the problem in a meaningful way.
The precautionary principle is often misinterpreted to mean: if in doubt, do not act. But its true meaning is: if in doubt 1) act to prevent possible harm and 2) act to reduce doubt. It reminds us that doubt creates an obligation to act, not an excuse to accept a situation, even a situation where lesser harms are done only to prevent greater harms. Because acceptance of mitigation as the end point of welfare science leads us down the path of accepting that these fundamental ethical problems are 'normal' or that it is morally acceptable to live with them rather than act to bring them to an end.