A new program called PULSE will test OpenAI and Anthropic AI models across 10 state, local, tribal, and territorial public health jurisdictions.
Key facts
- •The PULSE program will support trials in 10 state, local, tribal, or territorial jurisdictions.
- •OpenAI and Anthropic have donated 10 enterprise licenses capable of supporting up to 2,000 public health practitioners.
- •The program is scheduled to begin in autumn 2026, with final guidance expected in 2027.
- •Five primary use cases include biosurveillance, social determinants of health mapping, and automated clinical-data retrieval.
- •Nearly 40% of local health departments are currently not using AI, according to data cited by the Coalition for Health AI.
The Coalition for Health AI (CHAI) has launched the Public Health Use Case and Learning Scaling Engine (PULSE), a program designed to test generative AI tools within public health departments. Supported by OpenAI, Anthropic, and Accenture, the initiative provides 10 enterprise licenses for up to 2,000 practitioners. The program aims to develop implementation guidance for agencies considering future AI deployments, with pilot activities scheduled to begin in autumn 2026.
By the numbers
Program Scope and Use Cases
The program will focus on five specific use cases: biosurveillance and drug-wave prediction, social determinants of health mapping, operations and efficiency, public communications and multilingual translation, and automated clinical-data retrieval using the FHIR standard. CHAI’s leadership council will select the 10 participating jurisdictions and assign practitioners to these focus areas. Accenture is tasked with overseeing participant onboarding and assisting in the creation of implementation playbooks based on the trial results.
Data Governance and Operational Uncertainty
While the program aims to provide practical experience, CHAI has not yet defined specific technical, privacy, or security requirements for the pilots. The announcement does not clarify whether the models will process identifiable records, synthetic data, or de-identified information, nor does it specify how model outputs will be validated or reviewed by staff. Although OpenAI and Anthropic state that their enterprise services do not use inputs for model training by default, the program has not established rules regarding data retention, audit arrangements, or access controls for the participating agencies.
Timeline and Objectives
The pilot phase is set to commence in autumn 2026, with the final playbooks expected for release in 2027. These documents are intended to serve as reference material for other public health agencies. The initiative follows data from the National Association of County and City Health Officials indicating that nearly 40% of local health departments currently do not use AI, highlighting a potential gap in operational efficiency that the program seeks to address.
Advertisement
This article was independently rewritten by ManyPress editorial AI from reporting originally published by Artificial Intelligence News.
