Maryland Health AI Review Act

Overview

Maryland House Bill 820 (HB 820), officially titled “Health Insurance – Utilization Review – Use of Artificial Intelligence,” establishes a comprehensive regulatory framework for the application of artificial intelligence (AI), algorithms, and other software tools in health insurance utilization review processes within the state of Maryland. Enacted as Chapter 747, this legislation mandates that health insurance carriers, pharmacy benefits managers (PBMs), and private review agents adhere to specific guidelines when employing AI for evaluating the necessity and appropriateness of healthcare services. The core objective of HB 820 is to safeguard patients by ensuring that AI tools enhance, rather than replace, human clinical judgment and do not lead to unfair or discriminatory practices in healthcare access.

The bill explicitly prohibits AI tools from being used to deny, delay, or modify healthcare services, underscoring a commitment to patient-centered care and preventing technology from becoming a barrier to essential medical treatment. Instead, AI must be utilized in a manner that supports decisions based on an enrollee's individual medical history, clinical circumstances, and relevant medical records, rather than solely on group datasets. This focus on individualized care is a critical component, aiming to prevent generic algorithmic conclusions from overriding specific patient needs. Furthermore, the legislation requires robust oversight, including the development of written policies for AI use, regular performance reviews, and detailed reporting to the Maryland Insurance Commissioner, thereby establishing a clear accountability structure for the deployment of AI in this sensitive sector.

Definitions

Maryland HB 820 provides specific definitions to clarify the scope and application of the legislation. Central to the bill is the definition of “Artificial Intelligence,” which is described as a machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments. This broad definition ensures that a wide range of AI technologies, including various algorithms and software tools used in utilization review, fall under the purview of the new regulations.

The bill also defines the key entities responsible for compliance. A “Carrier” refers to any health insurance carrier that issues or delivers health benefit plans, policies, or certificates in the State of Maryland, thereby encompassing a broad spectrum of health insurers. “Pharmacy Benefits Manager” (PBM) is defined as an entity that manages prescription drug benefits on behalf of health insurers, self-insured employers, and other payers. Lastly, a “Private Review Agent” is an entity that performs utilization review for a carrier or PBM. These definitions are crucial for delineating the responsibilities and obligations of various stakeholders involved in the health insurance ecosystem and for ensuring that the regulatory requirements are applied consistently across all relevant parties utilizing AI in utilization review processes.

Governance and Institutional Framework

The governance and institutional framework established by Maryland HB 820 places significant responsibilities on health insurance carriers, pharmacy benefits managers (PBMs), and private review agents. These entities are mandated to ensure that any artificial intelligence, algorithm, or other software tool used for utilization review adheres to strict operational guidelines. A primary requirement is the development of comprehensive written policies outlining how AI tools will be integrated into the utilization review process, specifying the oversight mechanisms in place, and detailing how decisions informed by AI will be ultimately made by qualified human professionals. This emphasizes a human-in-the-loop approach, preventing AI from making autonomous final decisions on patient care.

Oversight is further strengthened by requiring these entities to conduct quarterly reviews of their AI tools' performance. These reviews are designed to assess the accuracy, fairness, and overall impact of the AI systems on utilization review outcomes, ensuring continuous monitoring and adjustment as needed. The Maryland Insurance Commissioner plays a central role in this framework, receiving quarterly reports from carriers detailing the use of AI in adverse decisions and grievances. The Commissioner is also tasked with compiling annual summary reports based on this data, providing a statewide overview of AI's impact on health insurance utilization review and identifying any systemic issues or areas requiring further regulatory attention. This multi-layered approach to governance aims to foster accountability and transparency in the deployment of AI within the healthcare insurance sector.

Key Focus Areas

Maryland HB 820's key focus areas are centered on ensuring that artificial intelligence in health insurance utilization review is used ethically, equitably, and in a manner that prioritizes patient well-being. A paramount provision is the strict prohibition against AI tools denying, delaying, or modifying healthcare services. This establishes a fundamental safeguard, ensuring that AI serves as a supportive tool rather than a decision-maker that could impede access to necessary medical care. The bill mandates that all utilization review determinations, even those informed by AI, must be based on an enrollee's individual medical history, current clinical circumstances, and relevant medical records. This prevents reliance solely on generalized group datasets, promoting personalized care decisions that consider the unique context of each patient.

Further critical aspects include the requirement that AI tools must not replace the decision-making authority of healthcare providers. This reinforces the principle that medical judgments should remain within the purview of qualified human professionals. The legislation also explicitly prohibits AI from discriminating unfairly against enrollees and mandates that AI tools be applied equitably across all populations. To ensure accountability and patient safety, carriers are required to protect patient data privacy when using AI and to implement measures that ensure these tools do not directly or indirectly cause harm to an enrollee. These provisions collectively aim to integrate AI into healthcare utilization review in a way that upholds patient rights, promotes fairness, and maintains the integrity of medical decision-making.

Implementation Framework

The implementation framework for Maryland HB 820 outlines a structured approach for carriers, pharmacy benefits managers (PBMs), and private review agents to integrate artificial intelligence responsibly into their utilization review processes. A cornerstone of this framework is the requirement for these entities to develop and maintain comprehensive written policies regarding the use of AI, algorithms, or other software tools. These policies must clearly articulate how AI will be employed, the specific parameters for its application, and the robust oversight mechanisms that will be in place to monitor its performance and ensure compliance with the bill's provisions. The emphasis on documented policies aims to create transparency and provide a clear operational guide for all personnel involved in AI-assisted utilization review.

Beyond policy development, the bill mandates ongoing performance monitoring. Entities utilizing AI in utilization review are required to conduct quarterly reviews of their AI tools. These reviews are critical for assessing the effectiveness, accuracy, and fairness of the AI systems, allowing for timely identification and rectification of any issues that may arise. The legislation also specifies an effective date of October 1, 2025, providing a transitional period for affected organizations to develop and implement the necessary policies, procedures, and technological adjustments to comply with the new requirements. This phased implementation approach is designed to facilitate a smooth transition and ensure that all stakeholders are adequately prepared to meet their obligations under the new law.

Monitoring and Evaluation

Monitoring and evaluation are integral components of Maryland HB 820, designed to ensure ongoing compliance and to assess the impact of artificial intelligence on health insurance utilization review. The bill mandates that health insurance carriers submit quarterly reports to the Maryland Insurance Commissioner. These reports must include detailed information regarding the number and outcome of all adverse decisions that involved the use of an artificial intelligence, algorithm, or other software tool. This reporting requirement is crucial for tracking the prevalence and influence of AI in decisions that directly affect patient access to care, providing a quantitative measure of AI's involvement in the utilization review process.

In addition to the carriers' individual reports, the Maryland Insurance Commissioner is tasked with a broader evaluative role. The Commissioner is required to compile annual summary reports based on the data collected from all carriers. These summary reports will provide a comprehensive statewide analysis of how AI is being utilized in utilization review, identifying trends, potential areas of concern, and the overall effectiveness of the regulatory framework. The Commissioner is further required to provide copies of these summary reports to the Governor and the General Assembly, ensuring that legislative and executive branches are regularly informed about the state of AI use in health insurance. This continuous feedback loop is essential for enabling ongoing policy adjustments and ensuring that the regulatory environment remains responsive to technological advancements and evolving healthcare needs.

Penalties, Liability, and Appeals

Maryland HB 820 integrates the regulation of artificial intelligence into the existing framework for health insurance utilization review, meaning that penalties, liability, and appeals processes generally align with established insurance laws. While the bill itself does not introduce entirely new penalties specifically for AI misuse, it mandates that AI tools must be used in a manner consistent with existing regulations governing utilization review. This implies that any non-compliance with the AI-specific provisions, such as using AI to deny, delay, or modify healthcare services, would be considered a violation within the broader context of utilization review regulations, potentially subjecting carriers, PBMs, or private review agents to existing enforcement actions by the Maryland Insurance Commissioner.

The bill's focus on adverse decisions and grievances, particularly the requirement for carriers to report on AI's involvement in such decisions, suggests that existing appeal mechanisms for patients remain applicable. If an AI-informed decision leads to an adverse outcome, enrollees would likely follow the established grievance and appeal procedures outlined in Maryland's insurance laws. These processes typically allow patients to challenge adverse decisions, with provisions for internal review by the carrier and external review by independent bodies or the Insurance Commissioner. The legislation aims to ensure that the introduction of AI does not diminish patient rights or access to redress, but rather integrates AI within a framework where accountability and the ability to appeal unfavorable decisions are maintained.

Relationship to Other Instruments

Maryland HB 820 operates by amending and adding to specific sections of the Maryland Annotated Code, Article – Insurance. Specifically, the legislation repeals and reenacts, with amendments, Section 15–10A–06 and adds a new Section 15–10B–05.1. This approach demonstrates that HB 820 is not a standalone piece of legislation but rather an enhancement and modernization of existing insurance regulations to address the emergence of artificial intelligence in healthcare. By integrating AI provisions directly into the established legal framework for utilization review, the bill ensures coherence and avoids creating parallel or conflicting regulatory structures.

The amendments to Section 15–10A–06 primarily concern the quarterly reporting requirements for carriers, now explicitly mandating the inclusion of information regarding the use of AI in adverse decisions. The addition of Section 15–10B–05.1 introduces the core requirements for the use of AI in utilization review, including the prohibitions, safeguards, and policy development mandates. This legislative strategy signifies that the Maryland General Assembly views AI in health insurance as an evolution of existing practices rather than an entirely novel domain requiring completely separate legislation. Consequently, the interpretation and enforcement of HB 820 will likely draw upon the established precedents and administrative practices associated with the broader Maryland Insurance Article.

National/Federal Alignment

Maryland HB 820 represents a proactive step by a U.S. state to regulate the use of artificial intelligence in a specific and critical sector: health insurance utilization review. While there is currently no comprehensive federal law in the United States specifically governing AI in healthcare or insurance, this state-level legislation contributes to a growing patchwork of state laws addressing various aspects of AI. It aligns with broader national discussions and emerging federal guidance that emphasize principles such as transparency, fairness, accountability, and the prevention of harm in AI systems. For instance, federal agencies like the National Institute of Standards and Technology (NIST) have developed AI Risk Management Frameworks, and the Office of the National Coordinator for Health Information Technology (ONC) has explored AI in healthcare, all advocating for responsible AI development and deployment.

This Maryland law can be seen as a practical implementation of some of these principles within a regulated industry. It sets a precedent for how states can tailor AI regulations to address specific industry concerns, particularly in areas where patient outcomes are directly impacted. While federal laws like HIPAA (Health Insurance Portability and Accountability Act) govern patient data privacy and security, HB 820 extends regulatory oversight to the application of AI to that data within utilization review, ensuring that the technology itself adheres to ethical and equitable standards. The existence of such robust state legislation may also influence future federal policy discussions by providing real-world examples of regulatory approaches and their effectiveness in practice, potentially serving as a model or contributing to a broader national standard for AI in healthcare.

Implementation Timeline

Compliance Checklist

Sources and References