Artificial intelligence (AI) is rapidly transforming the employment landscape, offering unprecedented efficiencies in recruitment, hiring, and talent management. However, with these advancements come significant regulatory challenges, particularly concerning fairness, transparency, and the prevention of algorithmic discrimination. Maryland stands at the forefront of this regulatory wave with its landmark Maryland AI Employment Law (HB 1106), set to become effective on October 1, 2025. This law mandates specific responsibilities for employers utilizing AI in employment decisions, with AI Impact Assessments forming the bedrock of compliance.

For compliance officers, general counsel, and CTOs at mid-to-large companies, understanding and implementing robust AI Impact Assessments is not just a best practice—it's a legal imperative with real consequences. This guide provides an authoritative, step-by-step approach to conducting these assessments, equipping your organization with the knowledge to navigate Maryland's requirements and leveraging an AI compliance software solution like AICompliant to streamline the process and ensure adherence to this critical legislation. Proactive implementation of an effective AI risk management framework is key to avoiding penalties and fostering ethical AI use.

Understanding the Maryland AI Employment Law (HB 1106)

The Maryland AI Employment Law (HB 1106) marks a significant development in US state-level AI regulation. This bill, effective October 1, 2025, targets employers who use "automated decision tools" (ADT) in employment-related decisions. Such tools include AI-powered systems used for:

Hiring and firing
Promotion and demotion
Compensation and benefits
Disciplinary actions
Job assignments
Performance evaluations

The law requires employers to perform specific analyses and provide disclosures when using ADTs. Critically, it aims to prevent algorithmic discrimination by mandating assessments of an ADT's impact on protected characteristics. Non-compliance with HB 1106 can result in significant penalties, up to $10,000 per violation, enforced by the Maryland Commissioner of Labor and Industry. This makes a clear understanding of your obligations and a structured approach to compliance essential.

The Imperative of AI Impact Assessments

An AI Impact Assessment (AIA) is a systematic process designed to identify, analyze, and mitigate potential risks and negative impacts associated with the development, deployment, or use of an AI system. Under Maryland's HB 1106, AIAs become a fundamental requirement for any employer leveraging AI in their HR processes.

What Constitutes an AI Impact Assessment?

For the purpose of HB 1106, an AIA should examine how an ADT may:

Produce disparate impacts or biases against protected classes.
Lack transparency in its decision-making process.
Pose risks to data privacy or security.
Potentially violate other employment laws.

Beyond Maryland, the concept of AIAs aligns with global best practices and emerging regulations. The EU AI Act (Regulation (EU) 2024/1689), with enforcement for high-risk systems beginning August 2, 2026, also emphasizes risk assessments for high-risk AI systems. Similarly, the Colorado AI Act (SB 24-205), effective June 30, 2026, will require developers and deployers of high-risk AI systems to exercise reasonable care to avoid algorithmic discrimination, with a strong focus on impact assessments. These converging requirements underscore the need for a comprehensive AI compliance platform that can manage diverse regulatory landscapes.

The objective is to not just identify risks but to demonstrate a proactive effort to mitigate them, ensuring the equitable and transparent application of AI in employment. This proactive stance is vital for algorithmic discrimination prevention and building trust in AI systems.

Step-by-Step Guide to Conducting Your Maryland AI Impact Assessment

Conducting a thorough and compliant AI Impact Assessment requires a structured approach. Here's a practical, step-by-step guide to help your organization meet the demands of Maryland's HB 1106.

Step 1: Scope Definition and System Identification

Begin by clearly defining the scope of your assessment. Identify all automated decision tools (ADTs) currently in use or planned for deployment that fall under HB 1106's purview for employment-related decisions. This might include:

Resume screening algorithms
Video interview analysis tools (like those regulated by the Illinois AI Video Interview Act (HB 2557), effective January 1, 2020)
Personality assessment tools
Performance management systems utilizing AI
AI-driven compensation analysis tools

For each identified ADT, document:

Its purpose and intended use.
The specific employment decisions it influences.
The vendor/developer information (if third-party).
Any existing documentation from the developer regarding fairness or bias testing.

This initial mapping is crucial. An AI compliance tool can help you log and categorize these systems, creating a foundational inventory for your assessments.

Step 2: Data Governance and Bias Identification

At the heart of algorithmic fairness is the data used to train and operate AI systems.

Data Collection & Sourcing: Document how data is collected, its sources, and any consent mechanisms. Scrutinize whether data collection practices inadvertently introduce bias.
Training Data Analysis: Analyze the characteristics of the data used to train the ADT. Are there demographic imbalances? For instance, if a system is trained predominantly on data from one demographic group, it may perform poorly or unfairly for others. Referencing principles from California AB 2013 (Training Data), effective January 1, 2025, which focuses on the responsible use of training data, offers valuable insights, even if not directly binding in Maryland.
Bias Detection: Utilize technical tools and methodologies to identify potential biases within the ADT's outputs. This could involve disparity analyses across protected characteristics (race, gender, age, disability, etc.) to see if the tool produces systematically different outcomes for different groups. For example, does a resume screener disproportionately rank candidates from certain demographics lower without a job-related justification?

This step is critical for algorithmic discrimination prevention. Tools like AICompliant's compliance checker (/tools/compliance-checker) can assist in evaluating the potential for bias and guiding this analysis.

Step 3: Risk Assessment and Mitigation Strategies

Once potential biases and risks are identified, the next step is to assess their severity and develop mitigation strategies.

Risk Evaluation: For each identified bias or risk, evaluate its potential impact. How severe could the discriminatory outcome be? What is the likelihood of it occurring? This aligns with the principles outlined in the NIST AI Risk Management Framework (AI RMF 1.0), which provides a comprehensive approach to managing AI risks, even though it is voluntary.
Mitigation Plan Development: Develop concrete, actionable strategies to mitigate identified risks. This might include:
- Data Diversification: Improving the diversity and representativeness of training data.
- Algorithm Adjustments: Working with developers to refine algorithms to reduce biased outcomes.
- Human Oversight: Implementing human review checkpoints for critical decisions made by the ADT.
- Transparency Mechanisms: Ensuring explanations for ADT decisions are clear and understandable to affected individuals.
- Regular Auditing: Scheduling periodic technical audits of the ADT for fairness and accuracy.
Accountability: Assign clear responsibilities for implementing mitigation measures.

Adopting a robust AI risk management framework ensures that identified issues are systematically addressed.

Step 4: Documentation and Record-Keeping

Maryland's HB 1106, like many other compliance frameworks such as NYC Local Law 144 (of 2021), effective July 5, 2023, places a heavy emphasis on documentation. Organizations must be able to demonstrate their compliance efforts.

Assessment Report: Create a formal report detailing the entire AI Impact Assessment process, including:
- The scope and methodology used.
- Findings regarding bias and risks.
- Mitigation strategies implemented.
- Dates of assessment and responsible parties.
Policies and Procedures: Develop internal policies and procedures for the ongoing use and monitoring of ADTs.
Disclosure Statements: Prepare clear and concise disclosure statements for employees and job applicants about the use of ADTs in employment decisions, as mandated by HB 1106.
Version Control: Maintain a clear version history of ADTs and their associated assessments, as changes to the system necessitate new reviews.

An AI compliance platform offers unparalleled capabilities for centralized documentation, version control, and automated record-keeping. AICompliant's dashboard (/dashboard) provides a comprehensive view of all your assessments and compliance artifacts, allowing for easy retrieval during audits.

Step 5: Regular Review and Updates

AI systems are not static; they evolve, and so do the risks associated with them. Your AI Impact Assessments must be an ongoing process.

Scheduled Reviews: Establish a schedule for regular reassessments of ADTs, especially if there are significant changes to the system, its data, or its intended use.
Performance Monitoring: Continuously monitor the ADT's performance in real-world scenarios to detect any emergent biases or unintended consequences.
Feedback Loop: Implement mechanisms to gather feedback from employees and applicants regarding their experience with ADTs, which can inform subsequent assessments.

This continuous monitoring is crucial for maintaining compliance and adapting to the evolving regulatory landscape, which includes diverse AI compliance requirements by state and international standards like ISO/IEC 42001:2023, effective December 18, 2023, an international standard for AI management systems.

Practical Tips for Effective AI Compliance in Maryland

Beyond the step-by-step assessment, consider these practical tips to bolster your overall AI compliance automation strategy.

Cross-Jurisdictional Awareness

While focusing on Maryland, remember that many companies operate across multiple states and even internationally. Other notable regulations include:

Colorado AI Act (SB 24-205): Effective June 30, 2026, this law similarly emphasizes algorithmic discrimination prevention and transparency for high-risk AI systems.
Connecticut AI and Data Privacy (SB 1103): Effective October 1, 2025, this bill expands on data privacy with AI considerations, imposing penalties up to $5,000 per violation.
NYC AEDT Law (Local Law 144 of 2021): Effective July 5, 2023, this law specifically regulates Automated Employment Decision Tools (AEDT) in NYC, requiring bias audits and public disclosures, with penalties up to $1,500 per violation per day.
EU AI Act (Regulation (EU) 2024/1689): Effective August 1, 2024, with high-risk enforcement by August 2, 2026, this comprehensive regulation carries significant penalties, up to $35,000,000 per violation.

An effective US state AI law tracker and global regulatory overview are vital. Your chosen AI compliance platform should ideally offer a unified view and an AI compliance checklist 2026 that adapts to various jurisdictional requirements, helping you avoid cumulative penalties from different regulators.

Leveraging AI Compliance Software

The complexity and volume of AI regulations make manual compliance efforts increasingly unfeasible. This is where an AI compliance platform like AICompliant becomes indispensable.

Automated Impact Assessment Templates: Utilize an AI impact assessment template within the platform to standardize your approach across different ADTs and ensure all regulatory requirements are addressed.
Centralized Documentation: Store all assessment reports, mitigation plans, and disclosure statements in a secure, easily accessible system.
Continuous Monitoring: Implement features that track changes in ADT performance, detect potential biases over time, and alert you to new regulatory updates.
Risk Prioritization: Leverage the platform's analytics to identify and prioritize the highest-risk AI systems and compliance gaps.
Audit Trails: Generate comprehensive audit trails demonstrating compliance efforts for regulators.

AICompliant’s solutions are designed to provide automated AI compliance, reducing the burden on your legal and technical teams and ensuring you stay ahead of the curve. You can access regulatory insights and compliance tools directly through our platform, including our specialized /tools/compliance-checker for quick assessments and a detailed /dashboard for comprehensive oversight.

Training and Internal Policies

No technology can replace human understanding and commitment. Ensure your teams—HR, legal, IT, and development—are thoroughly trained on HB 1106's requirements, the importance of AIAs, and your company's internal AI governance policies. Regularly update these policies as regulations evolve and new AI systems are adopted.

Conclusion

The Maryland AI Employment Law (HB 1106) presents a critical juncture for organizations leveraging AI in employment decisions. By proactively embracing the AI Impact Assessment process, your company can not only ensure compliance with effective date October 1, 2025, but also champion ethical AI deployment, prevent algorithmic discrimination, and build a more trustworthy and equitable workplace. Leveraging a specialized AI compliance software is no longer a luxury but a strategic necessity for navigating the complex and evolving landscape of AI regulation, ensuring your organization remains resilient and compliant.

Ready to Streamline Your AI Compliance?

Don't let the complexities of Maryland's HB 1106 or other emerging AI regulations hinder your innovation. AICompliant provides an end-to-end platform designed to automate your AI Impact Assessments, manage risks, and ensure continuous compliance across all jurisdictions. Explore how our solutions can safeguard your organization from regulatory penalties and foster responsible AI practices.

Learn more about AICompliant and view our pricing plans here.

Frequently Asked Questions

What is the effective date for the Maryland AI Employment Law (HB 1106)?

The Maryland AI Employment Law (HB 1106) is set to become effective on October 1, 2025. Employers utilizing automated decision tools in employment decisions must be compliant by this date.

What are the penalties for non-compliance with Maryland's HB 1106?

Non-compliance with the Maryland AI Employment Law (HB 1106) can result in significant penalties, up to $10,000 per violation, enforced by the Maryland Commissioner of Labor and Industry.

What types of AI systems are covered by the Maryland AI Employment Law?

The law covers "automated decision tools" (ADTs) used by employers in employment-related decisions. This includes AI systems influencing hiring, firing, promotion, compensation, disciplinary actions, job assignments, and performance evaluations.

How does the Maryland AI Employment Law relate to other US state AI laws?

Maryland's HB 1106 is part of a growing trend of state-level AI regulation in the US. It shares similarities with laws like NYC Local Law 144, effective July 5, 2023, and the upcoming Colorado AI Act (SB 24-205), effective June 30, 2026, by focusing on transparency, fairness, and the prevention of algorithmic discrimination in employment contexts. Companies operating in multiple jurisdictions need a comprehensive AI compliance platform to track and adhere to diverse requirements.

Can an AI compliance platform really help with AI Impact Assessments?

Yes, an AI compliance platform like AICompliant is designed to streamline and automate the AI Impact Assessment process. It provides templates, centralized documentation, risk management frameworks, and continuous monitoring tools, significantly reducing the manual burden and ensuring that all regulatory requirements, including those of Maryland's HB 1106, are met efficiently and effectively.

Maryland AI Employment Law: AI Impact Assessment Guide 2026