High Empirical but Low content Validity

Is it unethical for an employer to use a selection measure that has high empirical validity but lacks content validity? Explain.