How Public Health Experts Can Leverage AI at the State Level to Improve Patient Outcomes

Valuable Public Health Use Cases for AI

AI can analyze massive, varied data sets — from sources such as electronic health records, demographic data, medical sensors, health journals and even social media — much faster and far more accurately than a team of statisticians ever could. Through this analysis, the use cases for public health become nearly limitless. Some of the most immediate and practical applications at the state level include the following:

Disease Surveillance and Prediction

During the COVID-19 pandemic, governments around the world used AI models to predict outbreaks and track the spread of the virus, which helped them implement timely interventions. The Centers for Disease Control and Prevention, for instance, has been using AI to track COVID-19 since 2023. The pandemic was globally devastating, but the ability to forecast potential hotspots and quickly respond may have helped prevent even worse outcomes, both in terms of public health and broader social and financial impacts.

AI for disease surveillance can also help track emerging threats such as the H5 bird flu, as well as more familiar diseases. Every year, for example, scientists and government agencies rely on disease surveillance to decide which strains of influenza will be most threatening, and this information is used in the development of vaccines. Public health departments in many states are actively engaged in disease surveillance. By introducing AI capabilities, they can vastly improve their ability to track, forecast and hopefully prevent dangerous disease outbreaks and ultimately save lives.

Distribution of Public Health Information

During the pandemic, states such as South Carolina stood up entire contact centers to address their constituents’ questions and concerns. As a companion technology (or in some cases, as a replacement), cities and states can now deploy AI chatbots that can field some of these questions and act as a trustworthy source of information.

Jurisdictions can use chatbots and retrieval-augmented generation to make sure that everyone gets the same responses to the same queries, helping to maintain consistency and control over the information being shared. Deploying chatbots allows for interactive conversations, making it easy for users to engage and get the answers they need without waiting for human agents. It is even possible to provide proactive messaging alerts sent by a chatbot that can then field and answer questions via text.

DISCOVER: Here are 13 ways AI enhances healthcare operations, patient care and treatments.

Addressing Security, Privacy and Trust Concerns

Every technology has its risks, and for AI, the concern is around data privacy and security, as well as the trustworthiness of information and potential biases. Health data is highly sensitive, and agencies must maintain HIPAA compliance and other standards.

That said, healthcare organizations have stored and transmitted HIPAA-protected electronic data for decades, including through telehealth and mobile health applications. AI is the next iteration of that.

Just as you wouldn’t expose sensitive data to software that is not HIPAA-compliant, you also wouldn’t put healthcare data into noncompliant large language models (or give the public the opportunity to do so). HIPAA-compliant LLMs — such as Nabla, Amazon Web Services’ HealthScribe and Google Cloud’s Vertex AI Search feature — do exist. Furthermore, steps can be taken to anonymize data for use cases that might involve chatbots for public information queries, for instance.

As you investigate AI for public health use cases, there are three core considerations that will help you maintain HIPAA compliance and avoid spreading misinformation or acting on biases:

1. Leverage the right technology for the job, which means ensuring you have the budget and resources to implement those tools.
2. Determine whether you can afford to implement the necessary safeguards, especially when utilizing cloud environments or third-party services.
3. Train users extensively.

Finally, I can’t overstate the importance of trust in leveraging AI. Communities must have confidence in AI-powered public health initiatives, and that requires transparency around AI’s use as well as methodical data governance. Otherwise, organizations risk implementing the best technology with the best capabilities, only to lose the community’s trust.

There’s nothing easy or simple about change, and introducing AI into public health initiatives is no different. But the genie is out of the bottle, and it can’t be put back in. Jurisdictions must take steps toward using AI to aid public health initiatives, and do so in a safe and responsible way.

UP NEXT: Improve public health data exchange with the cloud.