4 ways HHS plans to help shape a national strategy for healthcare AI

HHS has thought through the ways AI can and should become an integral part of healthcare, human services and public health. Last Friday—possibly just days ahead of seating a new secretary—the agency released a detailed plan for getting there from here.

While the document carries no immediate regulatory weight, it comes as a “roadmap” by which stakeholders can know HHS’s mind on how to keep or make healthcare AI trustworthy, ethical and accessible across all socioeconomic strata (aka “equitable”).

“While AI could significantly improve many aspects of healthcare and human services, it also presents possible risks that could lead to adverse impacts or outcomes, such as algorithmic bias that may unintentionally reduce equity or breach protected information,” the document’s executive summary explains. “Responsible AI use should ensure equitable access and beneficence, safeguard protected information, involve appropriate consent where applicable, and ensure appropriate human oversight where needed.” More: 

‘Most notably, AI should be viewed as a tool to support and inform efforts rather than the sole answer to problems in the existing landscape.’

HHS hopes to translate its vision into action by pursuing four key goals, as follows. 

1. Catalyze health AI innovation and adoption. 

Improving AI adoption in medical research and discovery “could hinge on expanding use cases, encouraging AI in different disease areas and promoting AI-ready data standards,” HHS notes. Already, the agency points out, it has been directing funding and resources toward intramural and extramural research programs that “develop or leverage AI in medical research and discovery (see NIH’s Bridge2AI and ARPA-H’s Transforming Antibiotic R&D with GenAI to stop Emerging Threats project, or ‘TARGET’).” More: 

‘In the future, HHS plans to share data interoperability guidelines, engage the public, and continue prioritizing safe, responsible, and responsive AI in its funding of both intramural and extramural research programs.’

2. Promote trustworthy AI development and ethical and responsible use.

AI use in medical research and discovery could present biosecurity, privacy, bias and other risks, HHS states in the strategic plan. The agency touts platforms it has established to get ahead of the threats, such as NIH’s Science Collaborative for Health Disparities and Artificial Intelligence Bias Reduction [ScHARe]) and the Executive Office of the President’s National Biodefense Strategy. 

‘Going forward, HHS will share national guidelines specific to health AI, create sandboxes for industry collaboration and explore the use of AI for dynamic AI risk assessment.’

3. Democratize AI technologies and resources.

Working directly with the public and making critical data tooling and infrastructure more accessible to stakeholders with lower access to capital “could expand the opportunity to conduct AI-empowered research and discovery,” the plan reads. To support this goal, HHS is engaging communities (see NIH’s Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Research Diversity (AIM-AHEAD) and standardizing research data (see NIH’s Common Data Element Repository). 

‘HHS will continue to promote public/private partnerships, support multi-institutional research collaborations, and ensure access to needed data and data infrastructure.’

4. Cultivate AI-empowered workforces and organization cultures.

“To help ensure long-term successful and safe adoption of AI in medical research and discovery, AI talent pipelines and organizational working models may need to be bolstered,” HHS writes. The agency is “developing talent internally (see NIH’s Data and Technology Advancement National Service Scholar Program) and externally (see NIH’s Administrative Supplements for Workforce Development at the Interface of Information Sciences, AI/ML and Biomedical Sciences).”

‘HHS will continue to promote apprenticeship programs focused on AI in medical research and discovery activities to bolster talent pipelines and share guidelines for AI governance to help organizations foster robust AI-enabled cultures.’

Bearing in mind that HHS is soon to have a new head—whether Robert F. Kennedy Jr. or someone else—read the full plan and related materials here.