Public health agencies across the United States will test generative AI tools under a new program involving the Coalition for Health AI, OpenAI, Anthropic, and Accenture.
The public health use case and learning scaling engine, known as PULSE, supports clinical trials in 10 state, local, tribal, or territorial jurisdictions. This program aims to develop implementation guidance for public health agencies considering similar deployments.
OpenAI and Anthropic donated 10 enterprise licenses that can accommodate up to 2,000 public health workers. Accenture will oversee participant onboarding and help develop a playbook based on the trial.
The program will give public health workers access to enterprise AI products from OpenAI and Anthropic. CHAI does not specify the products, model versions, or configurations that will be used, nor does it explain how the two providers will be allocated throughout the pilot.
Former Texas Health Commissioner Dr. David Lakey said, “Any major technology change will succeed or fail based on trust, governance, and execution.” “PULSE supports government agencies in this effort and is specifically designed for practical implementation.”
5 public health use cases
The Coalition for Health AI’s Leadership Council will select participating jurisdictions and assign practitioners to communities focused on five use cases. These include biosurveillance and drug epidemic forecasting, social determinants of health and SDoH mapping, operations and efficiency and community feedback analysis, public communications and multilingual translation hubs, and automated clinical data search and FHIR query engines.
NIST’s AI Risk Management Framework recommends evaluating AI systems according to their intended use, operating environment, affected parties, and potential consequences. CHAI has not published separate evaluation, privacy, security, or human review requirements for the five PULSE use cases.
FHIR is an HL7 standard used to electronically exchange health information between compatible systems.
Although PULSE includes an automated clinical data search and FHIR query engine, this announcement does not define the role of generative AI in that workflow. It is not stated whether the model generates queries, retrieves records, summarizes returned information, or combines these features.
The announcement also does not explain whether staff will check for incorrect queries, incomplete searches, or unsupported summaries before the information is used.
Some of the proposed applications may include demographic, geographic, clinical, or population health information. CHAI did not say whether the pilots would use personally identifiable records, anonymized information, synthetic data, or aggregated data sets.
The U.S. Department of Health and Human Services requires organizations covered by HIPAA to protect certain electronic health information. Its cloud computing guidance states that regulated organizations and service providers must meet HIPAA requirements when cloud systems create, receive, maintain, or transmit electronic protected health information.
HIPAA does not apply to all participating organizations or workflows. Its application depends on the institution, the data involved, and the functions performed.
OpenAI states that input and output from business services, including ChatGPT Enterprise and its API, are not used to train or improve models by default. Anthropic also says that by default it does not use inputs and outputs from commercial products to train its models.
These provider policies do not define how your PULSE deployment is configured. This announcement does not establish rules governing retention periods, access controls, audit arrangements, data retention requirements, or submissions of protected health information.
“We believe AI must be useful, safe, and accessible to those tackling society’s most important challenges,” said Felipe Miron, Head of Government Market Entry at OpenAI. He added that the license is aimed at allowing public health agencies to evaluate the tool through a structured process.
Governance and evaluation remain undefined
The pilot is scheduled to begin in fall 2026. CHAI expects the resulting handbook to be released in 2027 and used as a reference by other public health agencies.
CHAI does not publish the metrics used to evaluate the pilots, nor does it explain whether each use case will be evaluated based on separate technical, operational, privacy, and safety criteria.
The announcement also does not explain how the model’s output will be reviewed. It is not stated whether staff must approve public communications generated, verify translations, verify clinical information obtained, or review biosurveillance or drug wave output before use.
NIST guidance recommends identifying AI functions that require human oversight and training users to understand the system’s performance and limitations. Its generated AI guidance also covers testing, validation, monitoring, documentation, privacy, and administrative oversight.
“Public health teams are being asked to do more with less, and AI can help with that, as long as it is deployed with caution and appropriate guardrails,” said Elizabeth Kelly, director of beneficial deployment at Anthropic. She said PULSE allows practitioners to test tools in their own environments with privacy, governance and responsible use measures built in from the start.
Nearly 40% of local health departments were not using AI, according to data from the National Association of County and City Health Officials cited by CHAI. The coalition said some departments are interested in reviewing workflows and improving operational efficiency.
PULSE offers enterprise licensing, onboarding, a peer community, and implementation playbooks. CHAI does not specify minimum staffing, infrastructure, interoperability, or cybersecurity requirements for participating jurisdictions.
Eligible participants include state and territorial health departments, county and local government agencies, tribal authorities, Indian health organizations, and metropolitan health departments.
PULSE plans to translate its findings from 10 jurisdictions into guidance for wider use. The announcement did not explain how the handbook would take into account differences in agency size, technology systems, legal responsibilities, staffing, procurement arrangements, etc.
The announcement also does not explain whether output from biomonitoring, pharmacokinetic prediction, or clinical data retrieval will be used only for testing, presented to staff for review, or incorporated into operational workflows.
PULSE forms part of the Coalition for Health AI’s broader work on healthcare AI governance standards. In May, the organization announced plans to develop guidance covering eight governance areas through workshops and working groups involving more than 150 healthcare AI professionals.
Since then, the coalition has begun publishing handbooks on organizations’ AI policies, governance structures, and internal resources. Additional guidance will also cover other areas of healthcare AI management.
Separately, CHAI has been working with the Joint Commission on a governance handbook in line with the voluntary responsible use certification of AI in healthcare. The PULSE announcement does not state that participating public health agencies will be evaluated based on their accreditation.
Dr. Brian Anderson, CEO of the Coalition for Health AI, said public health agencies entered the COVID-19 pandemic after years of limited investment in technology. He said the PULSE program is aimed at giving government agencies hands-on experience with AI before widespread implementation.
“We know that AI will change the way public health is delivered. The question is whether we do it thoughtfully,” said Dr. Ashish Jha, former White House COVID-19 response coordinator. He said the program will test which applications work and document the results for other government agencies.
SEE ALSO: Bunkerhill raises $55 million to scale agent AI across healthcare systems
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expos in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events such as Cyber Security & Cloud Expo. Click here for more information.
AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.
