Precautionary Principle

The principle that where there is a plausible risk of significant harm from an AI system, protective action should be taken even in the absence of complete scientific certainty.

When in doubt, be careful. If an AI might cause serious harm and you're not sure, err on the side of caution rather than forging ahead and hoping for the best.

The precautionary principle is a strategic governance stance. In areas of uncertainty, it is better to implement safeguards and relax them later than to move fast and deal with the consequences. Your governance framework should apply this where potential harms are significant.