Public health agencies in the US are preparing to test generative AI tools from OpenAI and Anthropic through a new pilot programme focused on practical government use cases.
The programme, called ‘PULSE’, will involve the Coalition for Health AI, OpenAI, Anthropic and Accenture. It is designed to help public health departments understand where AI can support everyday work, and where stronger governance is still needed.
What Is The PULSE Programme?
‘PULSE’ stands for the Public Health Use Case and Learning Scaling Engine. The pilot will support trials across 10 state, local, tribal or territorial public health jurisdictions.
The goal is not simply to give agencies access to AI tools. The programme is also expected to produce implementation guidance that other public health teams can use later.
OpenAI and Anthropic are providing enterprise licences that could support up to 2,000 public health practitioners. Accenture will help with onboarding and the development of practical playbooks from the trials.
What Public Health Teams May Test
The programme is expected to focus on several public health workflows. These may include biosurveillance, drug trend prediction, public communications, multilingual translation, community feedback analysis and clinical data retrieval.
Those are sensitive use cases because public health work often involves population data, demographic patterns, location information and sometimes clinical records. That makes privacy, accuracy and human review especially important.
Why AI Could Help Public Health
Many public health departments work with limited staff, older systems and heavy administrative demands. AI tools could help summarize reports, translate notices, organize community feedback or speed up routine data work.
For example, a health department may use AI to draft plain-language public notices, sort incoming feedback by topic or help staff search through large documents more quickly.
Used carefully, these tools could save time for teams that are already stretched.
The Governance Question
The main challenge is trust. Public health agencies cannot treat AI output as automatically correct, especially when decisions may affect communities, medical information or emergency communications.
Any rollout will need clear rules for privacy, staff review, data handling, audit logs, accuracy checks and who is responsible when AI output is wrong.
Enterprise policies from OpenAI and Anthropic may limit how customer data is used for model training by default, but the public health setting still needs its own controls. Agencies will need to know what data can be entered, who can see it, how long it is retained and when human approval is required.
Why This Matters
Public health is one of the areas where AI could be useful, but also risky. Faster translation, better reporting and quicker pattern detection could help agencies serve communities more effectively.
At the same time, poor implementation could create privacy problems, inaccurate summaries or public messages that need correction.
The Bottom Line
The PULSE pilot is an early test of how generative AI might fit into public health work. If the programme produces clear guidance, it could help agencies adopt AI more responsibly.
The key will be using AI as a support tool for trained staff, not as a replacement for public health judgment, accountability or community trust.