Maryland AI Hiring Bias Guidance Is Still Pending

Maryland Gov. Wes Moore announced an AI policy framework on September 22, 2026, directing the state’s civil rights and labor agencies to explain how existing discrimination laws apply when AI is used in workplace decisions. The announcement begins that work; it does not contain the promised employment guidance or create a standalone statewide AI hiring rule.

Bloomberg Law’s September 22 report also describes the employment-bias guidance as forthcoming and part of the broader framework. For employers and workers, the immediate answer is that Maryland’s existing employment protections remain in force. The agencies have yet to publish the more specific explanation of how they will apply those protections to AI-assisted decisions.

Existing employment law applies to AI-assisted decisions

The Maryland Commission on Civil Rights’ employment guidance says discrimination law covers recruiting, interviewing, hiring, work conditions, discipline and discharge. The commission says the general state employment discrimination law applies to employers with 15 or more employees, while unlawful harassment complaints can involve an employer with one or more employee. It also identifies protections against retaliation for complaining about discrimination or participating in an investigation.

Using software to filter applications or recommend candidates does not take the resulting employment decision outside those protections. A screening score may shape who receives an interview; a performance score may influence discipline. In either case, the relevant question under existing law remains whether the employer’s decision unlawfully discriminates. Maryland’s forthcoming guidance is meant to explain that application of the law, rather than switch the law on for the first time.

The workplace measures have different legal statuses

Maryland’s published AI principles place current state-government rules, promised agency guidance, proposed legislation and workforce initiatives in the same framework. Their status matters: a restriction on AI used by an executive agency is not automatically a restriction on every private employer in Maryland.

Law already in force: Existing employment discrimination protections cover the workplace decisions described above, including decisions informed by AI. They do not depend on publication of the new guidance.
State-government policy already in force: Maryland’s Responsible AI Policy applies to AI used by or on behalf of executive branch agencies. It prohibits specified uses, including emotion analysis, real-time and covert biometric identification, and social scoring. That policy’s stated scope is state executive branch use, not private employment generally.
Agency guidance still to come: The Commission on Civil Rights and Department of Labor are to issue coordinated guidance on existing discrimination law and workplace AI. The Department of Budget and Management and Department of Information Technology are also involved for state workers. The framework supplies neither a publication date nor a detailed new compliance test for private employers.
Broader safeguards proposed: The administration plans to work with legislators and stakeholders on human oversight of consequential personnel decisions, including discipline and termination. It also seeks limits on invasive monitoring, such as social scoring, unlawful surveillance, covert or real-time biometric identification, and tracking used to undermine worker organizing. The framework does not itself enact those broader prohibitions.
Worker participation and training: The administration proposes a standing committee to engage unions representing state workers and a separate gathering of unions, businesses and higher education institutions to develop workforce recommendations. It also describes training and transition efforts already underway. Those programs and proposed forums are distinct from enforceable employment-bias guidance.

The union proposals address different groups. The planned labor-management committee concerns state employees and their representatives. The proposed wider safeguards concern workers across Maryland and would require further policy work and, where legislation is necessary, action by lawmakers.

What employers can examine before guidance arrives

An employer using AI in hiring can identify each point where a tool changes an applicant’s prospects: filtering a résumé, ranking candidates, scoring an interview or recommending rejection. It can then examine the data and criteria behind that output, who can override it, and whether a person reviews the result before a decision. This is an interim assessment of exposure under existing law, not a checklist Maryland has issued.

The same review can compare outcomes across protected groups and distinguish a vendor’s general testing from evidence about the tool as configured for that employer. A vendor’s assurance alone cannot establish what happens to a particular applicant pool. For AI interview assessments, the relevant decision is whether a score affects who advances, not merely whether a person makes the final selection.

Employers can also separate applicant screening from systems that score performance, recommend discipline or monitor workers’ activity. Those uses affect different decisions and may raise questions about organizing as well as discrimination. Maryland has identified monitoring and consequential personnel decisions as concerns, but its proposed wider safeguards are not yet a general private-sector rule.

The next decision is the agencies’ guidance

The civil rights and labor agencies’ coordinated guidance remains the next document that could clarify how Maryland expects existing discrimination law to operate when AI influences workplace decisions. The framework leaves its timing and detailed approach open. Limits on broader worker monitoring and the proposed channels for union participation remain at separate policy or legislative stages.