The Coalition for Health AI (CHAI), in partnership with OpenAI, Anthropic, and Accenture, has launched a new programme called the Public Health Use Case and Learning Scaling Engine (PULSE) to test generative AI tools across US public health departments. The initiative will support trials in 10 state, local, tribal, or territorial jurisdictions, aiming to produce implementation guidance for broader public health AI adoption.
OpenAI and Anthropic have each donated 10 enterprise licences with capacity for up to 2,000 public health practitioners. Accenture will oversee participant onboarding and help develop playbooks based on the trial outcomes. CHAI has not yet disclosed the specific products, model versions, or configurations to be used, nor how the two providers will be assigned across the pilots.
What's New / Specs
PULSE focuses on five public health use cases:
- Biosurveillance and drug-wave prediction
- Social determinants of health (SDoH) mapping
- Operations efficiency and community-feedback analysis
- Public communications and multilingual translation hub
- Automated clinical-data retrieval and FHIR query engine
The programme does not define the role of generative AI in the clinical-data retrieval workflow—whether models generate queries, retrieve records, summarise information, or combine these functions. It also does not state whether incorrect queries or incomplete retrievals will be checked by staff. CHAI has not published separate evaluation, privacy, security, or human-review requirements for the five use cases, though it references NIST's AI Risk Management Framework.
Why It Matters
The pilots are scheduled to begin in autumn 2026, with playbooks expected in 2027. Nearly 40% of local health departments are not using AI, according to data cited by CHAI. The initiative could help standardise AI adoption in public health, but significant governance and evaluation details remain undefined. CHAI has not specified data retention policies, access controls, or whether identifiable health information will be used. HIPAA may apply to some but not all participating organisations, and the announcement does not clarify how protected health information will be handled.
Our Take
PULSE represents a meaningful step toward structured AI evaluation in public health, but the lack of published assessment criteria, privacy safeguards, and human-oversight requirements raises concerns. Without clear metrics for success and failure, the programme risks producing guidance that is too generic. The involvement of major AI providers and a systems integrator is positive, but real value will depend on transparent reporting of outcomes and limitations.
FAQ
What is the PULSE programme?
PULSE stands for Public Health Use Case and Learning Scaling Engine. It is a CHAI-led initiative that provides enterprise AI licences from OpenAI and Anthropic to 10 US public health jurisdictions for testing five specific use cases, with implementation support from Accenture.
What are the five use cases?
The five use cases are: biosurveillance and drug-wave prediction, social determinants of health mapping, operations efficiency and community-feedback analysis, public communications and multilingual translation, and automated clinical-data retrieval using a FHIR query engine.
How will data privacy be handled?
CHAI has not detailed retention periods, access controls, or whether identifiable data will be used. OpenAI and Anthropic state that inputs and outputs from commercial services are not used for model training by default, but the specific configuration for PULSE deployments has not been disclosed.
When will results be available?
The pilots start in autumn 2026, with implementation playbooks expected to be released in 2027 for reference by other public health agencies.
Who can participate?
Eligible participants include state and territorial health departments, county and municipal agencies, tribal authorities, Indian health organisations, and large city health departments.
