Beyond chatbots: generative AI as public health infrastructure and a new Digital Social Determinant of Health

In the recent past, generative artificial intelligence (AI) has gone from being a novelty to a powerful force that affects people's ways of seeking information, decision making, and engaging with institutions. In healthcare and public health, the focus has been on leveraging AI chatbots for delivering health information, patient support, and conversational assistance (Chow et al., 2024;Aldhafeeri et al., 2025). These have shown great potential in making things more accessible, engaging and facilitating digital health applications. The chatbot story, however, may all but mask a more fundamental shift in the way things are going. Generative AI is becoming more of a go-between for people and health information. Today, millions of people are consulting AI systems to interpret symptoms, assess health risks, gain insights into health conditions, access mental health resources, and understand healthcare options. People are using AI systems to help them interpret symptoms, evaluate health risks, understand health conditions, seek mental health resources, and understand healthcare options. This momentum is expected to continue, and generative AI is transforming from a digital instrument into a societal infrastructure that is impacting the health knowledge ecosystem as it is developed, delivered, and used. This shift is part of an overall trend in public health to acknowledge the influence of digital technologies on population health. The Social Determinants of Health (SDOH) have long been recognized as socioeconomic and environmental factors that influence health outcomes, such as education, income, housing, work, and social supports (Chidambaram et al., 2024). Recently, researchers and international public health bodies have suggested that the opportunity to use digital access, connectivity, digital literacy, and use digital health services are new factors shaping inequities in health. However, current discussions have mainly addressed the infrastructure and access to technology, but not the increasing role of smart systems that actively interpret, synthesize and personalize health information. This ever-changing environment presents a significant conceptual disconnect around the place of Generative AI in modern public health. In this opinion article, the authors argue that boiling generative AI down solely to an "app store of chatbots" leaves the transformative potential of generative AI understated when it comes to public health. Instead, generative AI should be seen as an emerging public health infrastructure that has the capacity to impact at scale on population health outcomes. Also, reliable health data and information provided by AI can become a new Digital Social Determinant of Health (DSDH). Given that education, income and housing impact health outcomes, access to a reliable AI system could have an impact on health literacy, health behaviors, healthcare navigation, health preparedness and health equity. In this article, we are not suggesting that having reliable access to Generative AI is a new determinant of health that sits outside of the public health domain, but rather that it is the next step in the evolution of digital determinants of health. This view extends SDOH frameworks with intelligent, AIenabled access to health knowledge, introducing a conceptual basis for analysis of how Generative AI can transform health equity, public health communication and population health resilience in the increasingly digital societies we live in. This is especially true of Digital Public Health, which goes beyond the scope of clinical settings to include prevention, population health promotion, community resilience, and health resource equity. This article aims to present the opportunities of generative AI to public health, its long-term impact on health equity, public health preparedness, and population well-being as both public health infrastructure and a potential Digital Social Determinant of Health.Traditionally, the dissemination of public health information has been done via health care providers, government agencies, schools, and mainstream media. The digital revolution turned this all upside down and made it possible to access health information online directly. Generative AI is a paradigm shift in the health information landscape, and could one of the most impactful digital health transformations of our time. Recently, there has been a substantial growth of AI-driven conversational systems in healthcare and various other industries (Khan et al., 2024). Modern-day healthcare Chatbots can now help users interpret their symptoms, get information about medicines, preventive care advice, health education, and mental health support (Balcombe, 2023). The use of these technologies has increased over time and become part of health-seeking practices. This shift has more than a technological purpose. AI systems are becoming the first point of contact for many for health related questions. This change has changed the way health information is retrieved and understood. Instead of simply being told about from institutional sources, people now speak to AI systems that can respond uniquely to them (Antonie et al., 2025). This new AI-enabled health landscape opens up new opportunities for public health involvement. AI systems can be used to deliver information in multiple languages, adapt to different levels of literacy, and offer round-the-clock information, irrespective of geographical location. These can augment the effectiveness of public health interventions dramatically. However, there are significant questions that need to be addressed with the increasing use of AI-driven interactions. Who will be in charge of the information produced by these systems? How do you know that it is accurate? What measures should be taken to stop the spread of misinformation? The main question now is: what will happen when access to reliable AI is not evenly distributed amongst populations? The quoted questions indicate that generative AI is not just a communication tool. Instead, it is now turning into a fundamental part of the modern health information landscape. AI-enabled health societies also have implications for the global health infodemic. In the event of a public health emergency, Generative AI can quickly create and share evidence-based information, thereby enhancing public health communication, and also help with health promotion. Meanwhile, these systems can also produce misleading, biased or false health information that can be widely disseminated. Thus, generative AI can be a force both for good and evil in health infodemics, underscoring the necessity to treat it as a public health infrastructure that needs to be governed and continually assessed.Traditional social determinants of health have been defined as income, education, employment, housing, and social support. All of these factors have an impact on individuals' health as they can access health care services and achieve good health (Sedkaoui & Benaichouba, 2026). Social determinants of health are defined by the World Health Organization (WHO, 2024) as the conditions in which people are born, grow, live, work and age, noting that these conditions are influenced by the "distribution of power, resources and opportunities in society. Likewise, the Center for Disease Control and Prevention (CDC) identifies the following as essential factors that influence health outcomes: economic stability, education, healthcare access, neighborhood environment, and social context. These frameworks show that health is not only shaped by the health care system but also by factors that are beyond the health care system, such as structural and societal factors that affect the opportunities for health. These well-defined viewpoints have been reinforced in recent public health publications, and it is now acknowledged that digital determinants of health (DDOH) are important risk factors for access to health information, health service use, and health inequities, including access to digital technologies, internet connectivity, digital literacy, and digital health services (Richardson et al., 2022). The Digital Social Determinant of Health (DSDH) is not intended to supplant current Social Determinant of Health (SDH) theory, but rather to be applied to the existing concepts. Education, income and social context remain important determinants in access and use of digital technologies. Trustworthy Generative AI will in turn depend on these factors to facilitate the way people access, understand, and use health information. Therefore, DSDH is not a separate pathway from SDOH, but a further one that helps to accentuate social inequalities and further harms health in digitally connected societies. As healthcare went digital, social determinants have been extended to include technological determinants that influence people's ability to interact with the health systems of the 21st century. Access to digital has become a key factor in health equity in recent years. In this opinion article, the authors suggest that generative AI could be the next step in this idea. In particular, having access to reliable AI systems could become a Digital Social Determinant of Health. Generative AI differs from previous digital determinants, which were mainly used for accessing health information online or accessing digital health services, by being able to interpret, synthesize, contextualise and personalise the health information through natural language interactions. Thus, while access to a reliable generative AI is not solely tied to internet connectivity or digital literacy, there are additional factors involved. It demonstrates a person's capacity to access understandable, contextually appropriate, evidence-based health information that fosters health literacy, health care navigation, and informed decision making. From this perspective, generative AI is an emergent aspect of current digital determinants and not a substitute for them. Users of sophisticated AI systems could gain from greater health literacy, more insight into preventive health practices, increased awareness of health risks, and more efficient interactions with the healthcare system. AI systems can help to streamline complex medical information and deliver tailored educational materials and support informed decision making (Rabbani et al., 2025). On the other hand, people without reliable access to AI technologies may be subject to new types of disadvantage. Access to AI-powered health resources might be hindered by limited digital literacy, insufficient technological infrastructure, and economic constraints. This means that entrenched inequalities in health may be exacerbated, if not reaffirmed. This view is in line with the general commentaries on digital inclusion in public health. Digital inequalities already impact on access to telemedicine, health information and digital interventions. If the benefits of generative AI are concentrated in technologically privileged populations, it could lead to an increase in inequalities. Also, the term "AI divide" should be seen as a part of the digital divide. The AI divide can be defined as inequalities in access to trustworthy artificial intelligence (AI) systems, AI literacy, computational resources, language inclusivity and the ability to critically analyze AI-generated information, in addition to the traditional digital divide, which mainly refers to the differences in internet connectivity and access to digital technologies. Taking on the role of a Digital Social Determinant of Health does not mean that AI should be seen as a substitute for the other Social Determinants of Health. Instead, it recognizes that digital technologies now have a growing impact on people's ways of accessing health knowledge, understanding health risks and interacting with health systems. The accessibility, digital literacy, equitable deployment, interoperability, and governance of AI are thus key aspects to be built into future health equity strategies that must be addressed by public health policies (World Health Organization, 2025). Generative AI serves as a bridge between human and health information. Generative AI is distinct from previous digital tools, which mainly provided access to information, by actively interpreting, synthesising, and personalising information that will impact health understanding, decision making and healthcare navigation. Its function thus goes beyond access to digital content and contributes to the opportunities for health itself. Therefore, the trustworthy access to Generative AI is a second-order digital determinant, building on existing digital determinants and literacies. People need to have the right access to the internet and skills to interact with it, before they can benefit from AI powered health support. Generative AI therefore should be understood as a "family of Digital Social Determinants of Health (DsdH) that builds on and continues to expand existing public health frameworks" (Yang, R. et al., 2026). Access to AI can thus be expected to increasingly become a Digital Social Determinant of Health through its impact on health literacy, health service engagement, and ability to make informed health choices.The Discussions today often see generative AI as a series of conversational tools. Though chatbot applications are still significant, this view does not capture the greater public health potential of the technology. Beyond individual interactions, public health systems need capabilities. Effective disease surveillance, emergency preparedness, risk communication, community health education and community engagement rely on good management and dissemination of information at the population level. It is now crucial to differentiate between Generative AI and predictive or analytical artificial intelligence. Traditional machine learning and predictive AI are used mostly for the analysis of structured data to reveal patterns, assess risk, predict disease outbreaks, and epidemiological surveillance. These models use statistical learning methods to make predictions, and are commonly used for outbreak detection, disease prediction and health risk modelling. Conversely, Generative AI is not intended for predictive modelling, but to understand, synthesize, generate, and communicate in natural language. Therefore, the benefits of Generative AI at the individual level should not be seen as a replacement for predictive AI, but as a complement to it within the overall framework of the public health system (Reddy, 2024). Predictive AI, for instance, can use surveillance data and predict the outbreak of a new infectious disease; while Generative AI can quickly compile and synthesize reports from various sources of surveillance information, interpret complex epidemiological findings, create multilingual public health advisories, create health education content for different audiences, and translate technical evidence into actionable policy briefs for decision makers. These additional features allow Generative AI to improve information created by traditional analytical systems to be more accessible, usable and shared. Generative AI has capabilities that are unique to them and could help with these functions at scale. In a recent study, Branda et al. (2025) identified the increasing importance of AI systems in public health emergencies. But, emergency response is just one application field. In addition to the role of emergency response, Generative AI can help public health agencies by consolidating, analyzing, and providing context to a variety of sources of unstructured data, such as surveillance data, scientific publications, policy documents, social media conversations, community feedback, and field reports. These systems can quickly create evidence summaries, detect emerging themes, enable multilingual communication and promote the timely dissemination of consistent public health messages using natural language synthesis. These types of capabilities are especially useful during a fast-changing public health event, when decision makers have to assess massive quantities of diverse data within short timeframes. When combined with existing population-level surveillance, health databases, communication platforms, and public health workflows, generative AI needs to be thought of as a Population Intelligence System that can augment population-level surveillance, communication, preparedness, and evidence-informed decision-making (Jahnel et al., 2024). The proposed Population Intelligence System should thus be considered a sociotechnical system, not a mere AI application. In this ecosystem, while predictive analytics and traditional machine learning produce epidemiological intelligence out of structured data, Generative AI enables better interpretation, knowledge synthesis, contextualisation, multilingual communication, and evidence translation. The value of human expertise is always critical in validating outputs and applying professional judgement and in using it to make decisions for the public good in an ethical manner. Importantly, the term "Population Intelligence System" does not mean that Generative AI carries out an autonomous disease surveillance or epidemiological forecasting. Instead, it is a digital universe that is integrated with predictive analytics that provide insights for the population, and Generative AI that helps interpret, contextualize, communicate, and translate knowledge. This holistic approach to Generative AI serves as a smart front-end for analytical results, seamlessly with public health public health and the general Population Intelligence System data sources, communication and decisions on this is crucial as the of Generative AI is to out analytical and deliver an contextually and knowledge for different not but knowledge synthesis, interpretation, and communication at scale. These systems could help improve public health surveillance of outbreaks, public health communication about health public health and public health response an outbreak using traditional surveillance systems, Generative AI could be used to create in multiple deliver scientific to health care on the education materials for respond to commonly public health and provide for policy makers. These improve the and of public health communication, but do not supplant the analytical of a surveillance system. This infrastructure approach is to that focus on of technologies (Reddy, 2024). AI applications are not to public health However, impact will only be if AI is to the public health system et al., 2024). should focus on the interoperability, and of AI into existing systems. the essential to have governance in place to the of predictive Generative AI, and human It essential that public health agencies Generative AI as a that the communication, and decision-making of public health, while for human and of evidence the public health decision-making Generative goes beyond simply to about enhancing the intelligence and resilience of as a public health have been of the impact of access to The of income, in education and access to health care remain as of in health The could with generative The digital divide has been a traditional of information and communication The public health may be a new that it must the AI the in access to and trustworthy AI systems. These can impact the access to AI-driven health information and decision support, a for People with AI technologies can get health education and decision making support. The following benefits may not be to This is especially important in resource skills are also for for people are and for people are digital and engagement have been shown to be crucial for the of digital health technologies in about digital health et al., 2026). If not and to public health can only Public health need to therefore make a to the AI divide. The to digital literacy, increase community engagement, infrastructure, and will be key to delivering equitable access to AI-powered public health AI to be in future public health digital literacy, and digital when a system of health. should therefore not only be seen as an ethical but as part of a understanding of how AI is integrated into public health systems to enable good public engagement, and system resilience over population health. is a key in the of public health interventions. Public in and sources of information is essential for health emergency response and disease This is the with generative are also of and in AI systems that have been identified 2024). are also in of AI such as the risks of risks, and governance et al., et al., 2026). This is especially significant as generative AI content that and its which and false information be a of public health systems. is a key of public health as with the infrastructure, AI infrastructure needs governance frameworks that can provide and All public health AI applications will need to have human as the critical AI be a substitute for public health systems, limited results, and surveillance should be expected in all public health for the use of AI (World Health Organization, 2024). If generative AI is to have the to with population level health care it must be in such a manner. Thus, must be seen as a for all public health systems and not just an ethical In the of in the and of AI systems, the potential benefits for the population health that generative AI will not and public health have been a for generative as by the of and discussions that have its generative AI has in the healthcare and public health with and discussions of the existing discussions have been on chatbot applications and conversational These technologies are however, they are just a part of the potential generative AI can make to public health. This is an opinion article that in its approach goes beyond the current on digital health and a conceptual It is not just a conversational it can also as an to facilitate knowledge synthesis, of multiple language communication, and informed decision-making within population health systems. The article also helps to the current public health landscape into a about health infrastructure, digital and health equity by Generative AI in that context. This opinion article related These are This will help to see the role of Generative AI, and understand the implications of this at the population level for health equity as a Digital Social Determinant of Health. These when offer a of the possible of Generative AI for the of public health systems and health opportunity in increasingly AI-enabled societies. Generative AI should be as a new public health infrastructure that can to surveillance, public health preparedness, communication, and engagement with the public and community & 2024). Importantly, this view does not suggest that an to traditional epidemiological modelling, predictive or public health it extends these capabilities by providing additional interpretation, communication and on the intelligence from existing analytical systems at the population level. access to reliable health information can become a digital social determinant of health that health literacy, health behaviors, health care navigation, and health equity. The of that the use of is considered a Digital Social Determinant of Health (DSDH) also builds on existing of digital health equity. access to AI systems, AI and could to be increasingly factors in shaping population health as societies become more on information This a basis for future of the impact of AI on health various and health care systems.