As we wrap up 2025, subject matter experts across SAS have been thinking ahead to 2026 and how we might see AI evolve in the public sector over the next 12 months.
Recently, we released a compilation of public sector AI predictions, including mine: AI-led extraction and entity resolution will continue to streamline public health reporting systems, revitalizing the foundation of our public health surveillance systems.
What do I mean by this? Simply put, over the next year, I see AI tools getting even better at extracting important information out of messy health care data, including figuring out which pieces of data belong to the same person, place or thing, even when they look different, making it easier for public health agencies to quickly and accurately detect emerging trends and threats.
What is the status quo and how will it evolve?
Today, AI already helps health systems extract and clean data, as well as match records for the same person/event/facility that are entered differently. However, it operates more as an intermittent assistant, helping at each stage of the process separately, rather than end-to-end, with humans checking each part of the process.
What prevents AI from being a more seamless tool? Right now, AI tools struggle with noisy, real-world clinical data. Sources are highly fragmented across different health systems and regions, and health systems often lack the infrastructure to deploy more advanced AI models at scale. Additionally, clinicians, public health professionals, and the general public are just beginning to become more comfortable with the idea of integrating AI into health care.
While all these challenges won’t be resolved overnight, in 2026, AI tools will be better at interpreting noisy data, making them much more useful and likely easier to deploy across different health systems. AI will more easily pull the important facts (e.g., categorizing the symptoms, diagnosis and location of a patient) out of sources that may be difficult to read, like handwritten notes. And, instead of providers and staff manually combing through records for matches to the same person in different ways (e.g., “Jon Smith,” “Jonathan Smith,” and “J. Smith”), AI will automatically recognize those entries as the same person.
How does this improve public health?
Here are a few examples of what this looks like in action.
- Faster outbreak detection: Currently, health agencies often receive data from hospitals, labs, and clinics in various formats, including paper and faxes, which can sometimes contain typos or missing details. AI could automatically pull out the important information (like symptoms, test results and patient locations) and piece together which reports may refer to the same case. This means public health officials might spot a spike in food poisoning or COVID-19-like symptoms days earlier than before – and a matter of days is crucial in public health.
- More immediate analysis during emergencies: During natural disasters, such as wildfires, or health outbreaks, like a surge in flu cases, hospitals and clinics experience a rapid influx of patients. AI could instantly gather and harmonize data from different hospitals and clinics, providing public health agencies with a real-time picture of what’s happening, such as which neighborhoods are most affected.
- Improved resource planning: Because of this improved ability to gather, clean and connect data in real-time, public health officials will more easily be able to see where supplies or other resources are needed most. For example, if public health agencies receive real-time data during and following a wildfire, they can more easily identify areas experiencing rising rates of asthma issues. It would also be much easier to identify trends such as which communities have low rates of vaccination.
- Better tracking over time: If someone goes to multiple health providers or locations, their data often ends up in separate systems. With improved entity resolution, AI could link those records correctly, helping public health officials follow disease trends more accurately and reduce duplicate data.
Looking ahead
While information sharing across departments and systems may remain a public health challenge, improved AI extraction and entity resolution will continue to streamline our reporting systems, supporting better surveillance systems and, therefore, better public health responses.
