US Public Health Agencies to Test OpenAI and Anthropic AI Tools Under New PULSE Program

Public health departments across the United States will test generative AI tools from OpenAI and Anthropic under a new initiative called PULSE -- the Public Health Use Case and Learning Scaling Engine. The program is led by the Coalition for Health AI (CHAI) in partnership with OpenAI, Anthropic, and Accenture.

PULSE will support trials across 10 state, local, tribal, or territorial jurisdictions, focusing on five use case areas: biosurveillance and drug-wave prediction, social determinants of health (SDoH) mapping, operations efficiency and community-feedback analysis, public communications and a multilingual translation hub, and automated clinical-data retrieval with a FHIR query engine.

OpenAI and Anthropic have each donated 10 enterprise licenses with total capacity for up to 2,000 public health practitioners. Accenture will oversee participant onboarding and help develop implementation playbooks based on the trial results. CHAI has not disclosed which specific products, model versions, or configurations will be used, nor how the two providers will be assigned across jurisdictions.

The CHAI leadership council will select participating jurisdictions and assign practitioners to communities organized around each use case. Pilots are scheduled to begin in autumn 2026, with resulting playbooks expected for release in 2027 as reference material for other public health agencies.

On privacy and governance, NIST's AI Risk Management Framework recommends evaluating AI systems by intended use, operating environment, affected parties, and potential consequences. However, CHAI has not published separate evaluation, privacy, security, or human-review requirements for the five PULSE use cases. The announcement also does not specify whether incorrect queries, incomplete retrievals, or unsupported summaries will be verified by staff before use.

Some proposed applications may involve demographic, geographic, clinical, or population-health information. CHAI has not specified whether pilots will use identifiable records, de-identified data, synthetic information, or aggregated datasets. OpenAI states that inputs and outputs from its business services are not used for model training by default, and Anthropic maintains a similar policy, but the announcement does not define retention periods, access controls, audit arrangements, or data-storage requirements for the PULSE deployments.

Data from the National Association of County and City Health Officials, cited by CHAI, shows that nearly 40% of local health departments are not yet using AI, with some expressing interest in revising workflows and improving operational efficiency.

The PULSE program represents one of the most significant attempts to bring generative AI into US public health operations at scale. Felipe Millon, OpenAI's head of government go-to-market, said the licenses are intended to help public health organizations evaluate the tools through a structured process. Elizabeth Kelly, Anthropic's head of beneficial deployments, noted that public health teams are being asked to do more with less and that AI can help when brought in with appropriate guardrails.