Good question. While they look they same they represent different things.

An arrow in a Bayesian network implies a statistical association, which goes both ways e.g. if A is associated with B then B is associated with A.

An arrow in a causal network, however, implies (one-way) causality e.g. A causes B but not the reverse