Employers may not use artificial intelligence in any employment decision — including recruitment, hiring, promotion, discharge, discipline, training selection, and terms or conditions of employment — if that AI has the effect of subjecting employees or applicants to discrimination based on any protected characteristic under the MHRA. This is a disparate impact standard: the employer need not intend to discriminate; discriminatory effect is sufficient. The protected characteristics include race, color, creed, religion, national origin, sex, gender identity, marital status, public assistance status, familial status, local commission membership, disability, sexual orientation, and age. Employers should conduct bias testing across these characteristics before deploying AI in employment contexts.

Employers must provide notice to employees and applicants when the employer is using AI in employment decisions covered by subdivision 9(b)(1) — i.e., recruitment, hiring, promotion, renewal, training selection, discharge, discipline, tenure, or terms and conditions of employment. Failure to provide this notice is itself an independent unfair employment practice. The bill does not specify the timing, format, or content of the notice, which creates compliance ambiguity — employers should err on the side of providing clear written notice before or at the time AI is used in the relevant employment process.