OpenAI and Anthropic's Role in PULSE Pilot

OpenAI and Anthropic join forces for PULSE pilot

OpenAI and Anthropic are joining a new public health program. It's designed to test generative tools in government workflows. This initiative, the Public Health Use Case and Learning Scaling Engine or PULSE, brings together a coalition of health organizations and technology partners to deliberately introduce high-level software into state, local, tribal, and territorial health jurisdictions. The program focuses on five specific use cases, ranging from biosurveillance and drug-wave prediction to multilingual communication hubs. So it's an important step.

Strategic placement in public sector infrastructure

The decision to deploy these tools across 10 jurisdictions signals a clear shift. But it's a move toward structured, evidence-based integration of advanced technology, and it doesn't allow for ad-hoc adoption. So the program gives 2,000 enterprise licenses to practitioners, creating a controlled environment for testing how large-scale models function within the constraints of public health operations.

That's a big step forward. They can't just wing it, though, so by involving entities like the Coalition for Health AI and external oversight partners, the architects of this pilot are aiming to establish standardized implementation playbooks for the future.

Evaluating the role of human oversight

Public health officials are balancing the promise of operational efficiency against the requirement for safety. But there's a major concern. The program aims to address questions regarding how automated systems handle data while supporting critical workflows such as clinical-data retrieval, and one key point of debate is how much human review is needed for outputs. It's a careful balance. While the program emphasizes the need for care and guardrails, the specific mechanics of validation for tasks like translation or biosurveillance remain a work in progress, so we can't call it finished yet.

Defining the scope of enterprise engagement

The program's technical backbone comes from OpenAI and Anthropic. Both providers have noted that, by default, their commercial enterprise services do not use customer inputs or outputs for model training, which addresses privacy concerns inherent in managing sensitive health data. But the program must still determine how these general policies align with the specific security and retention requirements of diverse public health agencies. It's a complex puzzle.

Every major technological transformation succeeds or fails based on trust, governance and execution. PULSE will support agencies in this endeavour, and is specifically designed for practical implementation. , Dr. David Lakey, former Texas health commissioner

Practical applications for public health

The pilot program outlines five distinct areas where these tools will be applied. These categories reflect common operational challenges in the public sector:

• Biosurveillance and drug-wave prediction models
• Social determinants of health and SDoH mapping
• Operations, efficiency, and community-feedback analysis
• Public communications and multilingual translation hubs
• Automated clinical-data retrieval and FHIR query engines

Moving toward standardized governance

The shift is clear. But nearly 40 percent of local health departments aren't using these tools yet, so there's ample room for expansion and a real opportunity to bring more agencies into the fold. The goal is to convert these initial findings into formal guidance by 2027, a timeline that reflects a desire to move from pilot testing to reliable reference material. Practitioners are expected to refine their internal policies as they gain experience with the provided technology. That's the plan.

Balancing speed with necessary caution

Public health teams are stretched thin. They're doing more with less, and the program's goal is simple: find applications that deliver real, tangible benefits without sacrificing one bit of accuracy. But Elizabeth Kelly, who leads beneficial deployments at Anthropic, says PULSE lets practitioners test tools in their own environments with privacy, governance, and responsible-use measures built in from the very start.

Preparing for the next phase of deployment

Testing matters. The transition from testing to operational use depends entirely on the quality of the findings that we generate in the coming months, and future efforts will focus on how agencies of varying sizes can adapt these playbooks to their specific legal and technical constraints. But the success of this pilot rests on the participants' ability to document their experiences effectively. As the sector moves toward the autumn 2026 start date, the industry will watch how these models hold up against real-world public health demands.

Frequently Asked Questions

What is the PULSE pilot program and who are its technology partners?

PULSE stands for Public Health Use Case and Learning Scaling Engine, a program designed to test generative tools in government workflows. OpenAI and Anthropic are the technology partners providing the program's technical backbone.

How do OpenAI and Anthropic address privacy concerns regarding sensitive health data in the pilot?

Both providers have noted that their commercial enterprise services do not use customer inputs or outputs for model training by default. This addresses privacy concerns inherent in managing sensitive health data, though the program must still align these policies with specific agency requirements.

What are the five specific use cases targeted by the PULSE pilot?

The five use cases are: biosurveillance and drug-wave prediction models, social determinants of health mapping, operations efficiency and community-feedback analysis, public communications and multilingual translation hubs, and automated clinical-data retrieval and FHIR query engines.

When is the pilot scheduled to start, and what is the timeline for formal guidance?

The pilot is set to start in autumn 2026, with the goal of converting initial findings into formal guidance by 2027. This timeline reflects a desire to move from pilot testing to reliable reference material.

Who is Elizabeth Kelly and what did she say about the PULSE pilot?

Elizabeth Kelly leads beneficial deployments at Anthropic. She said that PULSE lets practitioners test tools in their own environments with privacy, governance, and responsible-use measures built in from the very start.