Over the last decade, public health has gone digital. There are new systems for records, reporting, and analytics. But here’s the reality: tools like electronic health records, lab systems, immunization databases, emergency response platforms, and countless isolated spreadsheets often operate in isolation.
Are you noticing the pattern? Few systems are designed to easily share or reuse information. As a result, instant insights prove uncommon. Important data often gets trapped in silos, slowing down public health leaders and making it harder to collaborate during emergencies. Covid-19 made these gaps impossible to ignore. Even with plenty of data, sharing life-saving information quickly was a real challenge.
Ask yourself: Is the problem really about having too little data? No! It’s all about making everything work together and getting the details right so the information is actually useful. How frustrating is it to have the answers within reach, but be unable to access or combine them when you need them most?
So what comes next for public health data systems?
The answer is not to collect more data. The biggest challenge is making systems work together in ways that are actually useful in the real world. Public health teams need information they can trust when making decisions about outbreaks, staffing shortages, program effectiveness, risk, and community safety. AI and advanced analytics can absolutely support that work, but only when the underlying systems are reliable. That means uniform protocols, transparent data sources, strong privacy protections, repeatable workflows, and human monitoring built into the process from the beginning. Are we set to take that step and build a smarter, safer public health future together?
Public health systems were not built as one network
A common misconception in health technology is that digital systems become interoperable simply because they are electronic. Digitizing health records doesn’t guarantee smooth sharing. Data is only useful if it’s structured and governed for interoperability. Lab reports, case files, claims, and spreadsheets each speak a different language: making data integration a major challenge and slowing public health response.
Public health tech was built for varied needs, so every system operates differently. Experts spend lots of time cleaning up data, for example, fixing duplicates, standardizing terms, and aligning timelines. The most valuable AI tools simplify this messy work, not add flashy features.
AI is able to transform operations if used as infrastructure. It organizes messy fields, connects data, summarizes documents, and flags issues. These tasks turn raw data to actionable insight, speeding up responses during emergencies.
This infrastructure-focused approach is evident in MetaMation, which uses Microsoft AI Builder for quick no-code prototyping and expands functionality through CDC’s 1CDC Data Platform on Palantir Foundry. It supports AI-driven entity extraction and classification, natural-language search, fast querying, visualization, and automated ontology creation. It treats metadata as essential for public health intelligence. AI Builder enables no-code entity extraction, program classification, and natural-language queries. As it scales, MetaMation shifts to a governed platform that simplifies retrieval, visualization, ontology creation, and spreadsheet integration, making health metadata searchable, reusable, and well-governed. Automated metadata streamlines analytics and improves data reliability, while semantic infrastructure transforms digital data into useful insights, aiding effective decision-making.
The same principle is reflected in GENEVIC, an interactive tool powered by generative AI for visualizing and exploring genetic data., an interactive tool powered by generative AI for visualizing and exploring genetic data. Though GENEVIC focuses on genetics rather than public health, it demonstrates a careful approach to building reliable AI. Instead of generating random AI-generated results, GENEVIC’s HIPAA-compliant AI platform connects with organized databases and verified sources, cross-checks answers with programming outputs and peer-reviewed articles, and provides trustworthy information.
This strategy can inspire public health AI: systems should be built around reliable, well-governed data, clear workflows, and human monitoring, making AI part of a trusted infrastructure
AI in public health succeeds by enhancing operations and improving governance
Many people believe AI in public health is either miraculous or a threat, but its real value lies in assisting crucial public health workflows. Practical AI tools reduce delays, preserve crucial details, simplify tracking, and let experts concentrate on essential tasks. The main challenge is messy data: duplicates, confusing records, lengthy reports, and manual processes. AI can resolve these issues by organizing, summarizing, and highlighting patterns, permitting quicker and more effective decision-making in public health.
CDC’s enterprise-wide generative AI chatbot offers a concrete example of this friction-reduction model on a national scale. Built to support brainstorming, writing, coding, data analysis, document summarization, and program-specific chatbot creation, the tool redirected 41,460 staff hours to higher-value work, including analyzing 4,500 quarterly grant-recipient reports, saving an estimated 5,500 labor hours. Its impact shows that securely integrated, well-governed, and broadly adopted AI can greatly boost productivity when supported with the right workflows and training.
A key part of advancing public health AI is the CDC’s AI Strategy for FY 2026-2030 . Focusing on four pillars, accelerating adoption, strengthening governance and trust, developing data platforms, and building an AI-ready workforce, the strategy aims to reduce burdens, improve decision-making, secure platforms, establish standards, and implement risk-aware oversight. It shifts from isolated experiments to institutional capabilities, including agentic AI for responsive automation. CDC’s guidance stresses that while agentic AI can speed up research and evidence gathering, human monitoring and context continue to be essential for responsible public health decisions.
Better AI will not matter if the underlying systems stay broken.
AI alone can’t fix gaps in public health data infrastructure. Without strong governance and integration, even advanced models can worsen confusion. If models use incomplete or poor-quality data, their results may look certain but aren’t dependable. For critical public health decisions, this poses a major risk above technical issues.
Fundamental elements are critical: data standards, access controls, provenance, versioning, audit records, privacy, model assessment, and accountability. CDC’s Public Health Data Strategy outlines steps to improve communication among health organizations.
AI in public health also requires ethical guidelines: WHO guidance on large multimodal models emphasizes explainability, auditability, and population protection. The future of public health technology will likely depend less on individual AI models and more on whether institutions finally build infrastructure designed for connected intelligence instead of isolated reporting.
For years, the healthcare industry treated data collection as the main challenge. The next phase is about interpretation, coordination, and operational usability. The organizations that solve this problem well will be those that can turn fragmented information into faster, clearer, and more actionable decisions.
Author bio:
Anindita Nath is a Data Scientist with 10+ years of experience developing AI and machine learning solutions across public health, biomedical informatics, and large-scale data systems. Her work focuses on generative AI, LLMs, NLP, and data modernization, including leading pioneering agentic AI evaluation efforts in the federal public health sector and advancing AI adoption at the CDC.
Photo: Malte Mueller, Getty Images
