The Department of Health and Human Services (HHS) has released a nearly 200-page artificial intelligence plan, laying out its roadmap for how the department will oversee the emerging technology’s incorporation into the health sector.
Published on Jan 10., the plan aims to “set in motion a coordinated public-private approach to improving the quality, safety, efficiency, accessibility, equitability, and outcomes in health and human services through the innovative, safe, and responsible development and use of AI.”
HHS said its key goals with the plan are to: catalyze health AI innovation, promote trustworthy and ethical AI development and use, democratize AI technology, and cultivate workforces and organizations to safely use AI. The plan covers AI in medical research, products like drugs and devices, healthcare delivery, social services, and public health.
“At HHS, we are optimistic about the transformational potential of AI,” said Deputy Secretary Andrea Palm in a statement. “However, our optimism is tempered with a deep sense of responsibility. We need to ensure that Americans are safeguarded from risks. Deployment and adoption of AI should benefit the American people, and we must hold stakeholders across the ecosystem accountable to achieve this goal.”
One key focus of the plan is to improve the Federal government’s collaboration with “industry, academia, patients, and countless others,” acknowledging that a “whole-of-nation” approach is necessary for the successful and responsible implementation of AI in health and human services.
HHS plans to promote collaboration in several ways, including promoting partnerships between AI developers and National Institute of Health (NIH) funded investigators to accelerate the development of AI-powered medical products and treatments and expanding knowledge sharing between NIH-funded research institutions and under-sourced or underrepresented institutions.
Among increasing knowledge sharing, the department said it also wants to convene a public-private community of practice to share best practices for data-appropriate AI model use and develop a centralized repository of AI-ready data for authorized stakeholders.
The plan also highlights involving stakeholders – including the public and participants – in the research and development pipeline to gain more perspective on AI applications and opportunities. According to HHS, this could increase public engagement and education on the benefits, risks, and potential uses of AI.
Some of the department’s longer-term priorities focus on exploring privacy-enhancing technologies, developing policies for safe AI use outside controlled environments, and creating secure shared sandbox environments for collaborative innovation.
In addition to the long-term AI goals and visions of the department, several near-term priorities are covered in the plan such as expanding AI research funding across diverse disease areas and stages of research.
Others include prioritizing research collaborations across geographical borders, clarifying regulatory pathways for AI-enabled medical devices, issuing guidelines on AI risk management and governance for health organizations, establishing regional technical assistance centers to promote equitable access to AI, and exploring targeted training programs for AI governance and management in research and clinical settings.
Cybersecurity is another key focus of HHS’s plan, with the department noting that without risk management, AI may “put patient, participant, and public safety at risk, expose PII [personally identifiable information], and erode public trust in healthcare and public health systems.”
HHS said in its plan that it will address the current skilled workforce shortage through initiatives and training, standardizing best cybersecurity practices, simplifying cybersecurity implementation, and providing guidance on privacy and security trade-offs.
The department also said that it aims to integrate new cybersecurity requirements into its grants, contracts, and cooperative agreements. It will also develop guidelines on continuing operations after an AI system is compromised.
Cyberattacks against the health sector have been rising, with the Government Accountability Office (GAO) publishing a report in November calling on HHS to bolster the health sector’s cybersecurity. According to GAO, the department needs to do more work on security issues and implement its previous recommendations.
More recently, the department published a notice of proposed rulemaking in late December, stating that it wants to update the existing Health Insurance Portability and Accountability Act of 1996 (HIPAA) Security Rule with improved cybersecurity protections. In line with this, the AI plan aims to develop resources and educational materials to help organizations understand and address privacy concerns posed by AI.
The plan’s publishment follows the department’s increasing focus on AI after it reshuffled the organization to place its historically separated chief technology officer, chief data officer, and chief AI officer under a singular office. The three officials tapped to lead the new office were announced earlier this week.
