On October 30th, 2023, President Biden issued an Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence. The document largely outlines several federal initiatives and upcoming regulatory action, including establishment of NIST standards for AI Risk Management and reporting requirements for firms developing or selling AI capabilities. The executive order acknowledges the broad cross-industry impact and risk that AI presents, including in healthcare, energy, defense, housing, public safety, and workforce development. Themes include ensuring equity and inclusion, worker and consumer protections, reduction of biological and cyber geopolitical and warfare risk, and ensuring competition and innovation.
There are multiple components that we anticipate will directly impact the health and human services sector, including the following:
- Establishment of an HHS AI Task Force in early 2024 that will develop a strategic plan including regulatory action on responsible deployment and use of AI and AI-enabled technologies in HHS (including research and discovery, drug and device safety, healthcare delivery and financing, and public health). The scope of the AI task force includes:
- Healthcare delivery and financing, including quality measurement, performance improvement, program integrity, benefits administration, and patient experience
- Human oversight of AI-generated output
- Real-world performance monitoring and reporting on AI-enabled technologies
- Performance and monitoring requirements against discrimination and bias
- Incorporation of safety, privacy, and security standards for personally-identifiable information (PII)
- Dispersion of AI tools to State, local, Tribal, and territorial HHS agencies
- Reducing administrative burden and increasing workplace efficiency and satisfaction in HHS
- Provision of guidelines and assistance to payers and providers to clarify applicability of nondiscrimination and privacy laws to AI (April 2024)
- Development of common framework for HHS, Veterans Affairs, and other agencies to extend Patient Safety Organizations to capture AI clinical errors and other incidents that cause harm for centralized tracking (late 2024)
- Development of strategy for future drug development regulation (late 2024)
- Anticipated grant opportunities related to AI-enabled personalized medicine (personalize immune-response profiles for patients)
- Anticipated 2024 Leading Edge Acceleration Project cooperative agreement awards for initiatives that “explore ways to improve healthcare-data quality to support the responsible development of AI tools for clinical care, real-world evidence programs, population health, public health, and related research
- Anticipated grants related to the NIH Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity (AIM-AHEAD) program showcasing AIM-AHEAD activities in underserved communities
- Two Veterans Affairs nationwide AI Tech Sprint Competitions (anticipated late 2024)
- Anticipated HHS plan related to using AI systems for state and local benefit administration, including extensive assessment of fairness, effectiveness, and disclosure (April 2024)
The executive order also includes general provisions for multi-agency development of NIST standards to govern guidelines and best practices for safe, secure, and trustworthy AI systems. Much of the focus of these standards is related to national security threat mitigation, but in relation to healthcare initiatives, the NIST standards for audit and trust assurance will likely be applicable.
- The Coalition for Health AI (CHAI) released a statement of support shortly after the publication of the executive order. CHAI is a community of academic health systems, organizations, and expert practitioners of artificial intelligence (AI) and data science. CHAI also recently produced a Blueprint for Trustworthy AI.
- In response to a previous executive order on trustworthy AI (Executive Order 13960), HHS published a list of current AI use cases in the agency, as well as an HHS Trustworthy AI Blueprint (Sept 2021).