How can scientists and engineers systematically consider the societal consequences of their research at its early stages?