This claim's so big that it requires theoretical proof, empirical analysis isn't convincing (given the size of the claim). Causal inference experts have long known that many inputs map to outputs (that's why identification of the inputs that actually caused a given output is a never-ending task).