Labor & Employment Law Daily Wrap Up, AGENCY NEWS—HHS shares ‘OneHHS’ platform and other AI-driven ambits set to ‘transform agency operations’, (Dec 11, 2025)
By Leah S. Poniatowski, J.D.
HHS released its AI Strategy and a separate compliance plan, explaining the first steps the department will take to implement the current administration’s push to synthesize Ai technology into government operations.
HHS has been broadly tasked to update the department in light of several executive orders, especially EO 14179 removing Barriers to American Leadership in Artificial Intelligence, and Office of Management and Budget memoranda, specifically M-25-21 an M-25-22. In order to align with the ambits of those directives, HHS developed five pillars—found in HHS’ AI Strategy plan—to guide development and the “OneHHS” AI-infrastructure that will be used department-wide by all divisions, as briefly introduced in a December 4, 2025, HHS press release. HHS also stated that this phase is focused only on department operations.
Pillar 1: Ensure governance and risk management for public trust. In order to maintain public trust in the department as it further interweaves AI technology into its operations, HHS is modeling it internal guidance for risk management on the AI Risk Management Framework, as issued by the National Institute of Standards and Technology (NIST). Additionally, the department created an AI Governance Board with representation from senior leaders of the divisions under HHS that will meet biannually at a minimum and more frequently as needed.
The key strategic goals under this pillar are to establish “standardized minimum risk practices for high impact AI including pre deployment testing, AI impact assessments, independent review, monitoring, and safe termination if non compliant” in addition to “maintain an AI use case inventory, designating high impact AI, and publishing plain language summaries as appropriate.”
Pillar 2: Design infrastructure and platforms for user needs. With respect to design, HHS hopes to create a foundational suite of tools that can be centralized for security and use among the divisions so that new AI project teams do not need to start from scratch. Additionally, priority will be given to American-made technology and favor given to open-source tools and similarly transparent resources in compliance with applicable restrictions.
The key strategic goals under this pillar are to “emphasize centralization of data and governance to cut costs, reduce duplication, and enable secure, privacy-preserving, AI-ready access to HHS extramural and intramural studies, EHR, claims, cohorts, registries, surveillance, and human services data. Integrate governance and embed monitoring for cost effectiveness”
Pillar 3: Promote workforce development and burden reduction for efficiency. The vision HHS projects is a work environment where AI seamlessly enhances the worker’s capabilities, which will increase productivity and improve job satisfaction without compromise to quality, accuracy, and accountability.
The key strategy goals under this pillar include establishing “role based training pathways (introductory to advanced) aligned to mission roles (e.g., clinician, regulator, analyst, grants manager); [d]eploy secure, approved AI tools integrated with HHS systems and data, while systematically tracking key performance and security metrics; [s]tand up service desks and libraries for common tasks and prompts.” Additionally, HHS hopes to recruit and train top AI and data talent, including tapping public-private partnerships to quickly “upskill.”
Pillar 4: Foster health research and reproducibility through gold standard science. HHS also plans to use AI technology to discover and develop medications and other therapies, diagnostics, and clinical trial improvements that are “scientifically rigorous and openly accountable.”
The key strategy goals under this pillar include creating “centralized approaches for research and development (R&D) shared services for HHS-funded research where possible, while leveraging federated approaches where necessary.” At present, AI tools such as Google Colab and Anthropic’s AI for Science are being used to enhance the scientific process in work for HHS, and AlphaFold is being used for biological modeling. Additionally, HHS intends to develop AI performance assessment methodologies, data management and sharing tools for rapid discovery, and sharing resources across the government for support of large-scale computational needs.
Pillar 5. Enable care and public health delivery modernization for better outcomes. Finally, HHS has included the guidepost for AI use to provide “measurable improvements” broadly and individually to patients.
The key goals under this pillar are to identify critical public heath ailments that can be ameliorated with AI tools and promote AI-tool adoption. Additionally, HHS hopes to improve central coordination for its divisions in order to reduce duplication and to provide clear, unified messaging.
OneHHS. The strategy seeks to implement the “OneHHS” approach to bring all the HHS divisions in to collaborate on a single infrastructure that will “streamline workflows and enhance cybersecurity.” Relatedly, HHS identified “cross-cutting enablers” and the department is “undertaking new approaches to planning, surfacing, monitoring, and enabling future AI investments.”
Compliance Plan. In the separate Compliance Plan document, HHS explained the efforts it is undertaking to meet the requirements of the executive orders and OMB memoranda. HHS stated that it appointed and acting Chief AI Officer (CAIO) and established the governing board, as discussed in the accompanying strategy document.
Additionally, HHS provided target dates:
Development of a standard operating procedure to cover the AI use case waiver process: FY25 through FY26.
Application of minimum risk management practices to high-impact AI: April 3, 2026
Compliance of AI use cases currently used by HHS divisions: prior to April 3, 2026
Compliance with established risk management practices: ongoing
News: AgencyNews FederalRegulations AINews