What's Happening?
Maryland Governor Wes Moore has introduced a new framework for artificial intelligence (AI) governance, emphasizing the need to protect citizens, support workers, and ensure child safety. The framework addresses issues such as AI-generated intimate images,
impersonation, and algorithmic decision-making in insurance, including legislation to ensure qualified physicians, not algorithms, make final insurance coverage decisions. However, critics argue that the framework conspicuously omits a dedicated pillar for healthcare AI. Healthcare is identified as a critical system needing protection from AI risks, but this acknowledgment does not translate into a comprehensive governance structure for AI's use within the medical field. The rapid integration of AI into doctor's offices, hospitals, insurance systems, and homes, as well as its role in how patients seek health information and how medical decisions are supported, necessitates a specific and robust regulatory approach.
Why It's Important?
The absence of a dedicated healthcare pillar in Maryland's AI framework is significant due to the profound and unique implications of AI in medicine. While AI offers potential benefits like improved diagnosis, reduced administrative costs, identification of at-risk patients, and enhanced access to care, poorly governed systems pose substantial risks. These risks include the introduction of bias, compromise of patient privacy, generation of inaccurate clinical recommendations, and the shifting of accountability from human clinicians to algorithms without proper oversight. Healthcare represents a massive portion of Maryland's public finances, making effective AI governance in this sector an economic imperative. Without a tailored framework, the state risks failing to harness AI's benefits safely while exposing patients to potential harm and undermining trust in medical systems. The current approach, which relies on individual laws rather than a comprehensive strategy, may prove insufficient to manage the complexities of AI in healthcare.
What's Next?
Maryland should consider establishing a dedicated healthcare AI governance initiative. This initiative would involve a diverse group of stakeholders, including physicians, patients, technologists, ethicists, lawyers, health systems, insurers, and health occupations boards. Key areas for this initiative to address include determining which types of AI require validation before deployment, establishing clear guidelines for patient disclosures and consent, defining data protections for conversations with consumer AI systems, governing ambient recordings in clinical settings, ensuring clinician accountability for AI-assisted decisions, and creating mechanisms for patients to challenge algorithm-influenced decisions. The state needs to differentiate between AI used for administrative tasks and AI involved in clinical decision-making, as each presents distinct regulatory challenges. Given the governor's call for urgency and the recognition that federal guardrails are incomplete, Maryland has an opportunity to lead in responsible healthcare AI governance, leveraging its academic medical institutions and sophisticated healthcare ecosystem.
Beyond the Headlines
The deeper implications of Maryland's AI governance approach in healthcare extend to fundamental ethical, legal, and cultural shifts. The increasing reliance on AI for health information, as evidenced by a significant percentage of adults using AI chatbots for health queries, raises critical questions about data privacy and the nature of protected health information, especially when consumer chatbots may not be covered by HIPAA. The use of ambient AI scribes in patient encounters, while potentially reducing administrative burdens, transforms the doctor-patient relationship and necessitates clear policies on recording, storage, access, and the intimate nature of medical conversations. Without a comprehensive framework, there's a risk of AI shaping the healthcare system in ways that erode patient trust, compromise data security, and blur lines of accountability. Maryland's response to this challenge will not only impact its citizens but could also serve as a model or cautionary tale for other states grappling with the rapid advancement of AI in sensitive sectors.
