Harnessing artificial intelligence and digital technology for enhancing routine immunization among zero-dose children

Abstract

Objective

“Zero-dose children” remain a crucial burden to global health, particularly in low- and middle-income countries (LMICs) where vulnerable children have limited access to immunizations. This review explores the transformative potential of artificial intelligence (AI) in addressing this challenge and examines how AI can enhance routine childhood immunization.

Methodology

This narrative review synthesizes literature on AI, digital health innovations, vaccine delivery, and public health informatics. Literature was sourced from PubMed, Scopus, and Google Scholar, as well as institutional and authoritative reports such as the World Health Organization, the United Nations Children's Fund (UNICEF), and Gavi, the Vaccine Alliance.

Results

AI applications have demonstrated utility in identifying zero-dose populations, optimizing vaccine delivery, and supporting data-driven decision-making. Tools such as digital health passports, predictive analytics, and real-time monitoring platforms (e.g. UNICEF's real-time vaccination monitoring and analysis) enhance immunization tracking and reduce inequities. AI-driven chatbots, mobile applications, and social listening tools address vaccine hesitancy by providing tailored communication and combating misinformation. Furthermore, geographic information systems (GISs) and AI-based behavioral analysis improve outreach to underserved populations. However, challenges remain, including data quality, scalability, and ethical considerations.

Conclusion

AI solutions allow efficient identification of potential pockets of non-vaccinated populations, enhance decision-making processes based on data, and reduce vaccine skepticism due to AI-based interventions. Examples of successful implementation of AI include digital health passports, continuous monitoring mechanisms, and GISs for vaccine administration and coverage. This review also discusses the challenges and ethical considerations associated with AI implementation in LMICs, as well as the ethical implications. Using the insights derived from the review, this paper calls for the targeted utilization of AI toward the realization of a healthier future for children in need to achieve the target of reducing the number of zero-dose children by 50% by the year 2030, not only in areas afflicted by immunization disparities but across the globe.

Introduction

Zero-dose children—those who have not received any routine vaccinations, including the first dose of diphtheria-tetanus-pertussis—represent a significant challenge in global public health, particularly in low- and middle-income countries (LMICs) where access to immunization services remains limited.^1,2 These children are at significantly higher risk of morbidity and mortality.^3,4

Global estimates have shown a significant reduction in the number of zero-dose children, from 56.8 million in 1980 to 13.9 million in 2022. ⁴ However, despite this progress, in 2023, the number of zero-dose children increased to 14.5 million. ⁵ Roughly one-third of deaths among these zero-dose children are attributed to their unvaccinated status. ⁴ In 2020, 16.6 million of the 17 million zero-dose children were from LMICs. ⁶ Nearly 50% of zero-dose children are concentrated in urban areas, remote communities, and conflict settings. Identifying underserved populations—especially those involving zero-dose children and addressing vaccine hesitancy requires a targeted and strategic approach. COVID-19 has further exacerbated the situation, especially in LMICs, with the needs of zero-dose children now requiring an even greater sense of urgency. ⁷ Identifying populations that are not well-reached during vaccination campaigns and reducing hesitancy levels are two questions that any program focused on increasing zero-dose coverage must be able to answer.

Artificial intelligence (AI) is revolutionizing healthcare and other fast-evolving sectors; its emergence and growth have significantly shaped several aspects of medicine, including robotic surgery, biotechnology, aesthetics, pharmaceuticals, evidence-based medicine, advanced research, and transplantation techniques.^8,9 AI, by integrating extensively into medical developments, provides a holistic approach. Vaccinating zero-dose children is a difficult task that requires innovative and interdisciplinary approaches to achieve. The implementation of AI can contribute toward the realization of the global aim of reducing the number of zero-dose children by half before the year 2030. ¹⁰ By harnessing AI technology, public health initiatives can more effectively identify underserved populations, address vaccine hesitancy, and enhance routine immunization efforts. Embracing AI in these initiatives holds the promise of a healthier future for all children worldwide. This review highlights the transformative potential of AI in strengthening routine childhood immunization among zero-dose children in LMICs. It explores both specific AI applications, such as predictive analytics, digital health platforms, and behavioral tools, and their broader role in enhancing health system responsiveness, improving data use, and addressing equity in vaccine delivery.

Methodology

This narrative review explores the role of AI in enhancing routine immunization coverage among zero-dose children in LMICs. This review was structured as a narrative review due to the limited empirical research available on AI-based immunization strategies for zero-dose children. Our approach sought to identify conceptual linkages between AI and routine immunization coverage by drawing from related fields such as digital health innovations, vaccine delivery systems, and public health informatics. This allowed us to integrate evidence from both peer-reviewed literature and authoritative policy documents, building a broader perspective on how AI may contribute to reducing zero-dose children.

To ensure a comprehensive assessment, we employed a structured search strategy to identify relevant studies and policy documents. Literature was sourced from PubMed, Scopus, and Google Scholar, as well as institutional reports and authoritative sources such as the World Health Organization (WHO), United Nations Children's Fund (UNICEF), and Gavi, the Vaccine Alliance. The search was conducted using key terms, including “zero-dose children,” “AI in immunization,” “digital health interventions,” “vaccine equity,” and “machine learning in healthcare.” Studies were included if they met the following criteria:

1. Examined AI-driven or digital health interventions aimed at improving immunization coverage among zero-dose children.

2. Addressed challenges in vaccine delivery in LMICs.

3. Provided policy recommendations on the adoption of technology for public health and childhood immunization.

Results

The findings from the reviewed literature highlight multiple ways in which AI is being applied to address challenges in routine immunization, particularly among zero-dose children in LMICs. Evidence was synthesized thematically across five domains: data-driven targeting, behavioral analysis, targeted outreach, community engagement, and public–private partnerships, summarized in Figure 1.

Figure 1.

Harnessing opportunities in artificial intelligence (AI) and digital technologies to enhance routine immunization among zero-dose children.

Data-driven targeting: Identifying and engaging zero-dose populations

The AI algorithms can analyze extensive immunization datasets to identify trends, underserved areas, and influencing factors. This data-driven analysis can facilitate the recognition of trends, patterns, underserved regions, and influences of factors that encourage or hinder the progress of children's routine vaccination programs. ¹¹ With such information, health authorities can adapt and optimize program delivery with aligned resources. AI-driven data analysis should be accompanied by a data-driven process that establishes a sustainable cycle of improvement through iterative design and continues to campaign for uniform vaccine coverage.^12,13 The emergence of AI-enabled digital health passports represents a significant shift in vaccine-related record-keeping practices. ¹² These digital health passports, secured by blockchain technology and machine-learning algorithms, provide a tamper-proof method for maintaining vaccination records.^12,14 Additionally, health passports offer parents significant control through user-friendly mobile applications, allowing them to monitor and manage their children's immunization status, schedule appointments, and access other healthcare services. ¹⁵ In a validation study conducted at a children's hospital in Germany, parents using the VAccApp were able to complete vaccine-related questions accurately, suggesting that the app effectively aids in understanding and interpreting vaccination histories. The app was well-received, with users appreciating its ability to provide reminders for booster immunizations and keep them informed about their children's vaccination status.^15,16

UNICEF's support for the scale-up of digital health solutions, particularly real-time vaccination monitoring and analysis (RT-VaMA), is a significant initiative aimed at improving immunization coverage for zero-dose and under-vaccinated children in the East Asia and Pacific region. ¹⁷ RT-VaMA is a digital tool that facilitates real-time monitoring of vaccination data, allowing for the efficient collection and processing of information on vaccine coverage, utilization, and wastage. By leveraging Open Data Kit (ODK) technology, ¹⁸ RT-VaMA utilizes machine learning, including predictive analytics and natural language processing (NLP), trained on anonymized vaccination and demographic data collected ethically through partnerships with health authorities, adhering to WHO and UNICEF data protection guidelines.^19,20 Age grouping follows standard immunization schedules to tailor reminders effectively. ²¹ Additionally, real-time health monitoring tracks adherence and reports adverse events, allowing healthcare workers to address parental concerns and improve vaccine confidence through targeted interventions. ²²

The implementation of RT-VaMA provides several key benefits: timely data collection and analysis, enhanced immunization campaigns, reduction of inequities, and independence from internet connectivity. ¹⁸ RT-VaMA predictive analytics enable the identification of high-risk populations susceptible to vaccine hesitancy. By leveraging machine-learning models, AI can facilitate tailored interventions such as culturally sensitive educational content and interactive chatbot-based guidance, fostering informed decision-making. ²³ Additionally, AI-driven social listening tools can monitor digital platforms to detect emerging misinformation trends and counteract them with evidence-based health communication, thereby reducing the spread of vaccine-related misinformation. ²²

The RT-VaMA application also functions as an intelligent vaccine reminder system, notifying parents about the appropriate timing and type of vaccination. However, the ultimate decision to vaccinate remains with the parents, who may face barriers such as limited health literacy or social constraints. In this context, AI serves a crucial role in mitigating these challenges by analyzing individual behaviors, past adherence patterns, and demographic factors to deliver personalized vaccine-related messaging. For instance, AI-driven adaptive communication strategies can modify message tone, language, and content based on a parent's prior engagement with reminders, thereby enhancing responsiveness and adherence. ²⁴ Beyond reminders, RT-VaMA integrates behavioral reinforcement mechanisms, including positive reinforcement messages upon successful vaccination completion. ¹¹ The application can further enhance engagement by offering digital incentives, such as achievement badges, financial discounts, or public recognition within community networks, fostering a sense of social validation and motivation. ²⁵ While traditional reminder systems, such as Outlook calendar notifications, provide basic scheduling alerts, AI-enhanced platforms such as RT-VaMA offer advanced functionalities that significantly improve vaccination adherence. These include predictive analytics for identifying drop-off risks, personalized engagement strategies, and real-time sentiment analysis of public attitudes toward vaccination. ²⁴ Unlike static reminders, AI-driven systems dynamically adjust their interventions based on behavioral patterns, incorporating gamification and social incentives to optimize compliance rates. UNICEF piloted RT-VaMA in the Philippines during a polio outbreak in 2019, training 1500 health workers to use the ODK software for data collection on daily immunization rates, facilitating rapid decision-making at the national level. RT-VaMA proved instrumental in improving vaccine distribution and addressing stock level issues. ¹⁹ The positive outcomes in the Philippines prompted further adoption of the technology in other countries, such as Uganda and Nigeria, where it significantly increased vaccine coverage in rural and hard-to-reach communities. Building on these achievements, UNICEF is developing a comprehensive toolkit to help other countries implement RT-VaMA.^17–19 This initiative aims to empower governments to collect accurate, cost-effective data, thereby strengthening national and local health information systems. By scaling RT-VaMA, UNICEF hopes to enhance vaccination coverage globally, ultimately helping millions of children access the vaccinations they need.

The Expanded Programme on Immunization (EPI) in Pakistan employs various digital information systems to manage vaccination data and logistics. The Federal EPI Management Information System (MIS) is used in five provinces and territories to assign population targets, calculate coverage, and streamline vaccine-preventable disease reporting.^20,26 The Vaccine Logistics MIS, funded by the United States Agency for International Development, tracks vaccine distribution and consumption and is integrated with the Federal EPI MIS, but cannot easily be combined for comparative analysis. Zenysis Technologies, through its Sindh Sehat Analytics Platform (SSAP), aggregates data from fragmented sources to improve vaccine coverage and decrease zero-dose children in Pakistan.^27,28 Collaborating with EPI Sindh andemergency operations centers Sindh since 2019, SSAP uses data from over 25 sources, including the CBV network, Data Support Centre, EOC data team, and the Zindagi Mehfooz system, to perform detailed analyses, generate automated microplans, and link vaccinators with social mobilization workers. ²⁹

The application of AI goes beyond record-keeping by harnessing demographic data in the integration of information through a geographic information system (GIS), which equips the AI system with the ability to come up with a detailed map representing the vaccination coverage, population densities, and the distribution of healthcare infrastructures. ³⁰ The GIS mapping initiative facilitated by the Reveal platform shows promising intervention for identifying and reaching zero-dose children. ²⁶ Through local ownership, community engagement, expert support, and strategic partnerships, the initiative enhanced vaccination coverage and improved health service delivery. Addressing cost and technical expertise barriers through capacity building and strategic planning is essential for sustainable expansion. ³¹ AI, with its advanced capabilities, can overcome traditional barriers of reaching zero-dose children residing in remote, difficult-to-reach regions. AI can bridge the gap by identifying infrastructure, such as inadequate road networks, and determining possible transportation routes that could impact vaccine distribution. This data enables the health sector to easily plan actions to support these impacted communities. Such a panoramic view will help healthcare authorities identify areas with low levels of vaccination services and direct resources strategically to ensure equity in the provision of vaccination services.

AI for behavioral analysis and targeted interventions

Public health interventions attempt to target low-vaccination coverage groups, but balancing this with the challenge of outright vaccine hesitancy requires a unique solution. AI could be a tremendous enabler for a multidisciplinary response to these critical issues. Predictive analytics further empowers AI in the early identification of children likely to be missed during vaccination and in conducting early outreach to prevent outbreaks. ³² AI, through behavioral analysis, can inquire into possible reasons that are leading to vaccine hesitancy among caregivers of zero-dose children. ³³ The concerns expressed within social media messages, online forums, and other digital sources for surfacing the prevailing myths, misconceptions, and suspicions about vaccination can be analyzed through an AI algorithm.^13,34 Sentiment analysis of online discussions will bring to light areas that need targeted education campaigns to inform the choice of specific communication channels and messaging strategies that may work toward the encouragement of vaccination. The Cranky Uncle Vaccine Game, developed by the University of Cambridge's Social Decision-Making Lab, exemplifies the innovative use of AI in public health education. This interactive online game simulates conversations with vaccine-hesitant individuals, equipping users with effective strategies to address misinformation and promote vaccine acceptance. By engaging users in realistic scenarios and providing evidence-based information, the game enhances public understanding of vaccines and encourages informed decision-making. ²⁷ More concretely, AI-directed behavioral science interventions run with principles of behavioral economics in nudging the hesitant parent toward their child's vaccinations. ³⁵ It would code machine-learning algorithms to mine deep into the analysis of one's preferences, decision bias, and social networks to design tailored incentives and prompts customized to the parent's contexts. All of it is customized to parents’ contexts. Incentives help overcome psychological barriers and motivate parents to vaccinate their children through rewards, reminders, or social encouragement.

AI-driven targeted outreach: Reaching every child

AI has significant potential to revolutionize healthcare outreach, particularly in challenging geographical contexts. AI-powered telemedicine platforms enable virtual consultations, allowing healthcare providers to reach remote populations, ³⁶ thus increasing access to healthcare services without requiring physical presence. These platforms can dispatch timely reminders and educational materials to improve understanding of vaccinations and encourage adherence to schedules.^15,37 In Pakistan, the Zindagi Mehfooz (Safe Life) app, developed by Interactive Research and Development, digitizes child immunization records and sends automated short message service reminders to parents about upcoming vaccination appointments. This system has improved vaccination rates by ensuring parents are well-informed and reminded of their children's immunization schedules. ²⁹ Machine-learning algorithms offer real-time insights and decision-making support tools for health workers, helping them address vaccine hesitancy more effectively. AI-driven NLP chatbots provide virtual assistance in multiple languages, answering questions about vaccination schedules and side effects. ³⁸

AI-based chatbots have demonstrated groundbreaking potential to improve health outcomes in various medical fields, including mental health, oncology, pediatrics, and reproductive health. For instance, Woebot, a mental health chatbot, effectively delivered cognitive therapy to students with symptoms of anxiety and depression. Similarly, Tess, an AI-based behavioural chatbot, provided pediatric obesity and pre-diabetes treatment support to adolescent patients. AI-based chatbots have emerged as valuable tools in improving the immunization of children by providing reliable, accessible, and personalized health information.^39–42 One such example is Bablibot, which offers caregivers a bi-directional, 24/7 chat service at minimal cost. This service is especially critical in Pakistan, where traditional sources of immunization information include EPI cards, vaccinators at immunization centers, and call-based helplines. ⁴³ These chatbots offer confidential and remote support—crucial in regions with restrictive gender norms—ensuring all caregivers have access to accurate immunization information, thereby improving child immunization rates and public health outcomes.

Additionally, AI can coordinate mobile vaccination clinics or drone delivery to provide vital immunization services in remote areas, ensuring even the most inaccessible populations receive necessary vaccines. ⁴⁴ AI's integration into healthcare thus enhances worker capacity, optimizes vaccine supply, and expands access to health information, leading to more effective and widespread vaccination coverage. AI-powered mobile health (mHealth) applications send personalized reminders about upcoming vaccination doses, ensuring adherence to schedules and reducing the likelihood of children becoming zero-dose. ⁴⁵ These solutions facilitate real-time communication with families, simplifying the follow-up process with accuracy. AI systems provide tailored vaccine recommendations based on individual patient data, ensuring appropriate vaccination schedules.

AI-enabled platforms of involvement and trust-building in communities

This AI-driven community engagement empowers a significant step change in enabling a higher level of collaboration and communication among all engaged stakeholders to drive vaccination campaigns. ⁴⁶ This has been achieved by these platforms using NLP, which can capacitate stakeholders with effective engagement supported through a targeted intervention approach, effective campaigns, and dissemination of information on vaccination prospects. The use of AI also allows collaboration among public health agencies, the private sector, and works of philanthropy through the development of innovative technologies, data-sharing agreements, and sustainable financing mechanisms. ⁴⁷ Such AI operations can be scaled up with those vaccination programs so that, through this, the final goal of universal vaccine coverage would also be coupled with a robust immunization system.

AI techniques further upgraded the management of the supply chain of vaccines with respect to cold chain management by allowing risk forecasting and real-time alerts, and by preventing spoilage. ⁴⁸ The Jyotigram Yojana (JGY) is a rural electrification programme rolled out in India; it has contributed to improved availability of key child vaccines. Specifically, sustained cold chain storage for vaccines proves difficult in LMICs such as India because of high ambient temperatures and inconsistent electricity supply. JGY overcomes this by extending electrification to ensure efficient refrigeration for vaccine storage. ⁴⁹

Moreover, using information and communication technology—the electronic vaccine intelligence network system—has been implemented in several states in India. This system captures and tracks inventory, transportation, and storage conditions even down to the sub-district level. Therefore, it enables online tracking of cold chain points that assure the vaccines are still effective and can respond to the vaccination campaigns in the country. ⁵⁰ Going further toward the goals, AI-powered cross-sectoral collaboration programs integrate datasets across sectors to target an understanding of the social, economic, and environmental determinants of vaccine hesitancy and, therefore, set the stage for targeted interventions to bring changes in the underlying drivers of zero-dose populations. In other words, AI in community engagement platforms calls for importance when raising the uptake of vaccinations and reducing vaccine-preventable diseases across the globe.

Public–private partnerships: Collaborative solutions for immunization

Public–private partnerships, supported by AI, enable and harness opportunities to lower barriers to vaccine uptake. They bring together public health agencies with the private sector and philanthropic bodies to pool diverse expertise, resources, and networks that can advance the reach of the vaccination program. ⁴⁷ This is achieved through AI in the design of new and innovative technologies, data sharing, and the promotion of sustainable financing mechanisms—among many other AI-driven interventions—to build upon the necessary components that will unlock the full potential of universal vaccine coverage.

AI-driven pooling across public health bodies, academia, industry partners, and non-governmental organizations (NGOs) for a collective focus can also be pivoted on the challenges related to vaccine uptake. This would be possible since it would aggregate different datasets and take part in joint research programs, which would expose the social, economic, and environmental influential factors on vaccine hesitancy. ⁵¹ Public–private partnerships, cross-sectoral collaboration, and the use of AI represent a new way forward in supporting and advancing global immunization efforts and addressing the issue of zero-dose children.

Discussion

AI supports both data-driven decisions and personalized outreach. It has demonstrated substantial promise in optimizing immunization programs. However, alongside these advancements, critical challenges—such as ethical concerns, infrastructural limitations, data quality issues, and resistance to technology—must be addressed. This discussion explores the ethical, operational, and contextual considerations necessary for the successful implementation of AI-driven vaccination strategies, highlighting both success stories and ongoing barriers to scale and sustainability.

Ethical considerations in AI-driven vaccination strategies

Ethical AI governance frameworks guide stakeholders to be responsible regarding the design, implementation, and results of AI-driven immunization strategies. This is achieved through holding them accountable, ensuring transparency of the decision-makers between setting out the goals and methods articulated, and considering the possible risks that may arise with the AI technologies. ⁵² The second way accountability mechanisms allow stakeholders to work is through a priori consequences or ethical dilemmas. This is through the prior implementation. The preceding principles in ethical AI mean stakeholders are entitled to information available and relevant to how AI algorithms are developed, trained, and applied. ⁵³ This would give the whole AI intervention credibility toward the much-desired engagement with the affected community.

A central ethical priority is data security and privacy protection. AI systems for immunization rely heavily on sensitive health data, including children's vaccination records, demographic profiles, and in some cases, geolocation data. Without robust safeguards, there is a risk of unauthorized access, breaches, or misuse of personal health information. To mitigate these risks, AI governance frameworks advocate for compliance with international standards of data protection (e.g. WHO and UNICEF guidelines) and emphasize encryption, anonymization, and secure storage practices. Innovative approaches such as federated learning—where data remain on local devices while algorithms learn collaboratively—can reduce the risks of centralized data storage and protect privacy in low-resource settings. Additionally, role-based access controls and strict data-sharing agreements are essential to ensure that only authorized personnel can access sensitive immunization data.

Ethical AI governance frameworks will prioritize fairness, hence mitigating biases in data collection, algorithm design, and decision-making processes. It would be worth noting that a suggestion to make access to the underserved and hard-to-reach populations to vaccination services more equitable could also reduce health outcome disparities by suggesting AI technologies that factor in myriad perspectives and considerations of needs from the populace. Indeed, the ethical AI governance frameworks have consistently called for stakeholder engagement at all stages and steps in developing and implementing vaccination strategies. This is by the engagement and participation, so that there can be feedback on participatory decision-making to make the program more acceptable and efficient. This forms a part of meaningful engagement with the affected communities, healthcare providers, policymakers, and other stakeholders, considering that AI technologies must depict this diversity in their responsiveness to the needs and particulars of populations in different parts of the world. ⁵⁴

AI in action: Success stories of zero-dose population vaccination

A personalized AI-based smartphone app was implemented to enhance childhood immunization coverage in Pakistan. This app sends timely reminders to parents about vaccination schedules, provides educational content, and offers support through AI-driven chatbots. A randomized controlled trial demonstrated a significant increase in vaccination uptake among children aged 10 and 14 weeks compared to standard care. ³⁷

In Nigeria, the AI-Driven Vaccination Intervention Optimiser (ADVISER) was utilized to improve immunization strategies. ⁵⁵ The pilot study in Tanzania, Dar es Salaam, was conducted by the Clinton Health Access Initiative together with the Ministry of Health, focusing on urban areas to establish zero-dose children and vaccinate them. They employed AI methodologies to assess immunization data, assess constraints, and design tailored outreach initiatives. This approach managed to safely reconnect more than 1100 children who had not received the first dose or were inadequately immunized with the health system. ⁵⁶ India's Intensified Mission Indradhanush, launched in 2018 as part of India's Universal Immunization Programme, aimed to increase vaccination coverage to 90% across the country. It utilized AI to identify and reach zero-dose children, especially in urban poor and migrant communities. Health workers, supported by AI-driven data, conducted home visits to address caregivers’ concerns and ensure children received necessary vaccines (UNICEF). ⁵⁷ One notable example is the MyChild Solution, developed by UNICEF and Arm, which utilizes mobile technology to create a digital immunization registry. ⁵⁸ The implementation of this system has significantly reduced the number of children missing vaccinations by ensuring timely reminders and accurate tracking of immunization status.

Another successful initiative is the Connected Health AI Network (CHAIN) by macro-eyes, which has dramatically reduced vaccine wastage by 96% in regions of Tanzania. CHAIN leverages real-time data from frontline health workers to forecast vaccine demand accurately, ensuring that the right number of vaccines is available at the right time. This not only improves vaccine availability but also minimizes wastage, thereby optimizing immunization efforts. ⁵⁹

Challenges and barriers in implementing AI for zero-dose reduction

AI and other technologies are expected to enhance routine. However, several constraints and challenges have ensued that limit the efficacy and adoption of these technologies in this vital field of public health. While preliminary findings from pilot implementations such as RT-VaMA and ADVISER indicate promising outcomes, these results are often constrained by small sample sizes and context-specific limitations. Larger-scale, longitudinal studies are needed to validate the broader applicability and sustainability of these interventions across diverse LMIC contexts.

Some of the potential problems of AI systems are connected with the data inputs AI uses, as the performance depends on the quality of the entered data. The most common factor affecting immunization coverage in developing countries is that the recorded information can be incomplete, erroneous, or dated due to limited resources. A systematic review highlighted that data quality issues, such as inconsistency and incompleteness, significantly impact the effectiveness of digital health interventions, including AI applications.^53,60 A more effective strategy could be investing in hybrid models that combine AI with human oversight—where trained data stewards regularly validate AI-generated insights, which has the potential to enhance accuracy.^61,62 Additionally, most studies discussing AI in immunization ignore the role of community health workers (CHWs) in improving data accuracy. CHWs, who work at the grassroots level of immunization of children, can provide real-time, context-specific corrections to immunization records. Digital tools such as AI-assisted mobile apps for CHWs have been piloted in some regions (e.g. India's U-WIN platform) but require further scalability studies. ⁶²

Another challenge identified in studies is related to the use of modern technologies in teaching and management, which entails the support of certain key resources, namely the internet, electricity, and equipment. There are numerous areas with high numbers of zero-dose children, but this essential infrastructure is missing, which complicates the deployment of AI and other digital interventions. Furthermore, studies often discuss infrastructure barriers in broad terms but fail to differentiate between urban and rural implementation challenges. Urban centers in LMICs often have digital infrastructure but suffer from bureaucratic inefficiencies, whereas rural areas lack both resources and trained personnel. Tailored AI strategies should acknowledge these differences and incorporate adaptive implementation models. ⁶³ However, the assumption that AI requires high-tech environments has been challenged by emerging research on offline-compatible and low-power AI solutions. For example, edge AI—where algorithms run locally on mobile devices without needing cloud connectivity—has been proposed as a viable solution for rural immunization programs. ⁶⁴ One potential solution is leveraging AI-as-a-service models, where governments and NGOs pay for AI capabilities on demand rather than investing in costly in-house development. This approach has been successfully applied in other sectors (e.g. agriculture and supply chain management), but its feasibility in public health immunization remains underexplored. ⁶⁵

Several studies note that healthcare workers and policymakers in many LMICs lack sufficient training to use AI technologies effectively. ⁶⁴ However, they often recommend broad training programs without considering that workforce education varies widely across different healthcare systems. A one-size-fits-all training model is unlikely to be effective. Instead, a tiered approach—where AI literacy training is tailored to different user levels (e.g. CHWs, district health officers, and policymakers)—could enhance adoption rates. ⁶⁶

Additionally, resistance to AI is often attributed to fear of job loss among healthcare workers. ⁶⁷ However, this perspective may oversimplify the issue. In reality, healthcare workers may resist AI not only due to employment concerns but also because of a lack of trust in AI-driven recommendations. Studies on AI acceptance in LMIC healthcare settings suggest that interventions should focus on AI–human collaboration rather than AI replacement, emphasizing how AI can enhance rather than override clinical judgment. ⁶⁸

In addition, the gathering and utilization of health records caused a massive provocation of ethical and privacy concerns. It is crucial to make sure that data is properly collected and managed in a way that does not violate individual privacy and legal requirements, though it can be very demanding, particularly in environments where there are fewer rules in place that need to be followed. ⁶⁹ However, most discussions focus on regulatory compliance rather than practical implementation. In settings where formal data protection laws are lacking, decentralized AI models—such as federated learning, which processes data locally rather than transmitting it to central servers—could mitigate risks while still allowing AI models to learn from diverse datasets. ⁶¹

The use of AI technologies involves the costs of the development of these technologies, the deployment of these technologies, and the maintenance of these technologies. ⁷⁰ In most LMICs, available health financing remains weak, and there might not be adequate resources to devote to the adoption of such superior innovative technology solutions in preference to primary health requirements. ⁷¹ Implementing change and introducing new technologies in any field faces opposition from stakeholders who are comfortable with the old ways of working. This resistance can be due to ignorance, reluctance because of perceived loss of jobs, or the tendency to resist change due to the normal mode of operation. AI systems and other technologies that apply them have to interface and communicate with other electronic health systems. ⁷² Cultural beliefs and practices also affect the acceptance and outcomes of technological interventions. ⁷³ Expressions of technophobia or anti-vaccine attitudes in some communities can serve as barriers to achieving the immunization goals that AI seeks to support. A study on AI-based predictive models for vaccine delivery emphasizes the importance of considering cultural factors to enhance the effectiveness of AI interventions in immunization programs. ⁷⁴ A promising approach is embedding AI-generated recommendations within community-based decision-making frameworks. Studies from behavioral economics suggest that AI acceptance increases when technology is introduced through trusted local intermediaries. ⁷⁵ More research is needed to explore culturally adapted AI interventions tailored to specific regions and belief systems. The challenges, opportunities, and barriers associated with AI-driven vaccination strategies are summarized in Figure 2.

Figure 2.

Key challenges and barriers in implementing artificial intelligence (AI) for zero-dose reduction.

Conclusion

AI has the potential to transform how universal immunization coverage is achieved. It supports the fusion of analytical insights, programmable outreach, and innovative technology to ensure no child is left behind. This synergy paves the way for a future where all children are protected from vaccine-preventable diseases. AI, by improving both vaccine distribution and demand forecasting, will support equity in vaccination service delivery while benefiting from sustainable vaccine supplies. In addition, universal vaccination and child health are better supported with newer insights based on data using AI technologies and innovative public health tools.

Analyzing social media and digital platforms helps uncover misinformation and tailor communication strategies to build trust and promote vaccine acceptance among communities. Innovative AI technologies—such as electronic health passports and mobile applications—empower caregivers with real-time vaccination information and personalized reminders. These tools streamline record-keeping, enhance caregiver engagement, and ensure adherence to immunization schedules, thereby bolstering overall vaccination coverage and child health outcomes. Furthermore, AI fosters collaboration across sectors through public–private partnerships, driving innovation in vaccine development, distribution, and delivery systems. By integrating diverse expertise and resources, these partnerships accelerate progress toward universal vaccination goals while ensuring the sustainability and scalability of immunization programs.

By leveraging AI's transformative potential, we can envision a future where every child is protected from vaccine-preventable diseases, supported by innovative technologies and data-driven strategies that optimize health outcomes worldwide.