Skip Navigation
Skip to contents

JPMPH : Journal of Preventive Medicine and Public Health

OPEN ACCESS
SEARCH
Search

Articles

Page Path
HOME > J Prev Med Public Health > Volume 59(5); 2026 > Article
Perspective
When Deepfakes Meet Herd Immunity: Artificial Intelligence, Vaccine Avoidance, and the Next Infodemic Threat
Minsoo Jung1,2orcid
Journal of Preventive Medicine and Public Health 2026;59(5):555-559.
DOI: https://doi.org/10.3961/jpmph.26.549
Published online: August 27, 2026
  • 1,200 Views
  • 34 Download

1Department of Health Science, Dongduk Women’s University, Seoul, Korea

2Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA

Corresponding author: Minsoo Jung, Department of Health Science, Dongduk Women’s University, 60 Hwarang-ro 13-gil, Seongbuk-gu, Seoul 02748, Korea, E-mail: mins.jung@gmail.com
• Received: June 6, 2026   • Revised: July 24, 2026   • Accepted: August 10, 2026

Copyright © 2026 The Korean Society for Preventive Medicine

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

prev
  • Vaccination remains among the most effective public health interventions, but its impact increasingly depends on public trust and a reliable information environment. During the coronavirus disease 2019 (COVID-19) pandemic, misinformation became a major obstacle to vaccine uptake and disease control. The rapid development of generative artificial intelligence (AI)—including synthetic text generators, conversational chatbots, deepfake images, synthetic audio, and AI-generated videos—has changed how misinformation can be produced, personalized, and disseminated. This Perspective argues that AI-generated misinformation may represent a new stage in the evolution of the infodemic. Compared with earlier forms of vaccine misinformation, AI-generated content can be produced automatically, tailored to individual audiences, made visually persuasive, and distributed at unprecedented speed and scale. Drawing on historical vaccine controversies, health communication research, and epidemiological principles of herd immunity, we propose that AI-driven misinformation should be examined not only as a communication problem but also as a potential epidemiological risk factor that may weaken population immunity. The implications extend beyond vaccine hesitancy to the preparedness and resilience of public health systems during future pandemics. We discuss how anti-vaccination movements may change in the AI era and propose responses centered on trust-building, digital health literacy, information surveillance, and community engagement to support vaccine confidence and pandemic preparedness.
Vaccination has reshaped public health by preventing millions of deaths and controlling diseases that previously caused widespread morbidity and mortality. Vaccine acceptance, however, has never been universal. Since compulsory vaccination was introduced in the 19th century, anti-vaccination movements have drawn on recurring concerns about individual liberty, vaccine safety, government authority, and distrust of scientific institutions.
Misinformation is now a major contributor to vaccine hesitancy and vaccine avoidance. Controversies involving the measles-mumps-rubella vaccine, thimerosal-containing vaccines, and alleged vaccine injuries show how perceived scientific uncertainty can affect vaccination behavior even after specific claims have been refuted [1–3]. The coronavirus disease 2019 (COVID-19) pandemic further showed that misinformation can weaken public health responses across countries and health systems.
Generative artificial intelligence (AI) presents a distinct challenge for the public health information environment [2,4]. AI systems can produce synthetic content that closely resembles authentic health communication. Evidence specifically examining AI-generated vaccine misinformation remains limited; however, emerging studies suggest that generative AI can increase the scale, realism, and perceived authenticity of health misinformation, and broader evidence shows that online misinformation influences vaccine attitudes and health-related decision-making [2–4]. Deepfake technologies can fabricate visual content and imitate trusted experts, allowing misinformation to be created rapidly, repeatedly, and at scale. Despite these capabilities, direct effects on vaccination behavior and population immunity have not yet been established.
Current research has largely examined intermediate cognitive and behavioral outcomes rather than population-level epidemiological effects. In a systematic review, Saeidnia et al. [4] reported that most studies assessed trust in health information, perceived credibility, misinformation recognition, vaccination intentions, and health-related decision-making after exposure to AI-generated or digitally manipulated content. Downstream outcomes, including vaccination coverage and herd immunity, have rarely been studied. Thus, although the available evidence suggests that AI-generated misinformation may affect determinants of vaccination behavior, direct evidence linking such misinformation to reduced vaccination coverage or impaired herd immunity remains lacking.
Building on this evidence and these gaps, we propose a testable conceptual framework in which AI-generated misinformation increases exposure to false content through digital platforms, alters trust in vaccines and perceptions of vaccine-related risk, contributes to vaccine avoidance among susceptible individuals, reduces vaccination coverage, and ultimately increases vulnerability to vaccine-preventable disease outbreaks. This pathway should be read as a theoretically informed framework rather than an established causal model. Its purpose is to define priorities for future epidemiological research on how information disorders may affect population immunity.
The distinction between vaccine hesitancy and vaccine avoidance has become increasingly important. Vaccine hesitancy refers to delayed acceptance or uncertainty, whereas vaccine avoidance denotes a decision not to vaccinate despite vaccine availability. Historically, hesitancy has been shaped by media framing, anecdotal reports, and vaccine-critical information environments [5,6]. Social media intensified these influences by accelerating the spread of misinformation [7]. During the COVID-19 pandemic, vaccine misinformation also became more closely intertwined with political polarization, conspiracy theories, and declining institutional trust.
Generative AI may intensify these vulnerabilities by producing fabricated images, videos, and testimonials that appear highly realistic, although evidence showing corresponding effects on vaccination behavior remains limited. AI systems can also tailor misinformation to cultural, political, religious, or health-related contexts, making dissemination both scalable and targeted. These features may enlarge the reach, precision, and perceived credibility of anti-vaccination communication (Table 1). Whether these technological features ultimately reduce vaccination coverage remains an empirical question.
The potential public health significance of AI-generated misinformation can be evaluated through an epidemiological framework. For infectious diseases, the herd immunity threshold is defined as:
H=1-(1/R0)
where R0 is the basic reproduction number. For a disease with R0=10, at least 90% of the population must be immune to interrupt sustained transmission. In a hypothetical population of 1 million individuals with 95% vaccination coverage, and assuming high vaccine effectiveness, durable immunity, and homogeneous mixing, this level of coverage would be expected to sustain herd immunity.
As a conceptual illustration, suppose vaccination coverage declined by 10 percentage points after widespread exposure to AI-generated misinformation. This assumption is intended only to illustrate possible epidemiological implications and should not be interpreted as an empirical estimate of the effect of AI-generated misinformation. Coverage would decrease from 95% to 85%, and the unvaccinated population would increase from 50 000 to 150 000. Although simplified, this example shows how modest reductions in vaccination coverage may substantially increase outbreak vulnerability. In practice, herd immunity depends on multiple factors, including vaccine effectiveness, duration of immunity, population heterogeneity, geographic or social clustering of unvaccinated individuals, and contact patterns.
For highly contagious diseases such as measles, even small reductions in vaccination coverage can increase outbreak vulnerability. If AI-generated misinformation affects vaccination behavior at scale, its effects could extend beyond individual decision-making to transmission dynamics. We therefore propose that AI-generated misinformation be conceptualized as an emerging information-environment risk factor for population immunity, not only as a communication challenge. As vaccination behavior becomes increasingly shaped by the reliability of the information environment, misinformation may also influence outbreak potential. For example, a deepfake falsely depicting a senior public health official acknowledging severe vaccine injuries could appear credible and affect vaccine-related decision-making.
Research consistently shows that misinformation spreads online more rapidly than accurate information [8]. Generative AI can increase both the volume and realism of synthetic content. Automated dissemination systems and platform recommendation algorithms may then amplify that content beyond the capacity of conventional fact-checking.
AI-generated misinformation may reinforce cognitive biases by presenting synthetic content as authentic. It may also influence adherence to public health measures beyond vaccination during future emergencies. Such misinformation could reduce trust in public health agencies, promote ineffective treatments, and delay protective behaviors. Because effective pandemic responses depend on risk communication, institutional trust, and coherent public messaging [9,10], widespread AI-generated misinformation may weaken several conditions needed for outbreak control. Addressing these risks will require coordinated responses from public health authorities, technology platforms, healthcare professionals, community organizations, and international agencies (Table 2).
Traditional public health communication remains necessary, but it is unlikely to be sufficient when synthetic content is emotionally engaging, visually persuasive, and amplified by algorithms. Three priorities require coordinated action across public health, healthcare, technology, and community sectors.
First, vaccination campaigns should place trust-building at their center by working with community leaders, healthcare professionals, and local public health authorities as credible messengers. Second, digital health literacy should extend beyond general health education to include practical skills for assessing source credibility, recognizing manipulated content, and understanding how AI-generated misinformation is produced and circulated. Third, cross-sector collaboration is needed to build information surveillance systems that function analogously to infectious disease surveillance. These efforts should align with World Health Organization infodemic management approaches, including Information Network for Epidemics, while remaining adaptable to local contexts. Real-time monitoring of misinformation, when combined with digital epidemiology and risk communication, may support earlier interventions. Governance safeguards must protect privacy, transparency, and freedom of expression.
The COVID-19 pandemic showed that information disorders can weaken public health responses. Generative AI may intensify this problem by enabling realistic misinformation to be produced and disseminated at large scale. Future pandemics may therefore involve overlapping biological and informational crises. Protecting population health will require medical countermeasures as well as communication, surveillance, and governance strategies that preserve trust, strengthen digital resilience, and protect the public information environment. This Perspective proposes a testable framework that positions AI-generated misinformation as a potential risk factor for population immunity and identifies a need for empirical research and interdisciplinary preparedness.

Ethics Statement

This study is a perspective article based on secondary sources and does not require ethical approval.

Conflict of Interest

The author has no conflicts of interest associated with the material presented in this paper.

Funding

None.

Acknowledgements

None.

Author Contributions

All work was done by Jung M.

Table 1
Comparison of traditional and artificial intelligence (AI)-enabled vaccine misinformation across content generation, dissemination, and potential public health implications
Characteristics Traditional vaccine misinformation AI-enabled vaccine misinformation
Content creation Created manually by individuals or organized groups Synthetic text, conversational chatbot content, synthetic audio, and deepfake images or videos can be generated at scale with minimal human effort
Production cost and speed Constrained by human resources, technical expertise, and production time Produced rapidly with substantially lower technical and resource requirements
Dissemination Spread primarily through interpersonal communication and conventional social media sharing Dissemination may be amplified by coordinated bot networks and platform recommendation algorithms, which are distinct from AI-based content generation
Audience targeting Broad messaging with limited personalization AI systems can generate narratives tailored to cultural, political, religious, or health-related contexts
Perceived authenticity Depends largely on conventional text or multimedia formats May be increased by synthetic text, audio, images, and videos that appear authentic
Potential public health implication Localized effects on vaccine confidence and uptake Possible reduction in effective population immunity if AI-enabled misinformation contributes to widespread vaccine avoidance

AI-enabled misinformation includes both generative AI technologies (e.g., synthetic text, conversational chatbots, synthetic audio, and deepfake images or videos) and automated dissemination mechanisms (e.g., coordinated bot networks and recommendation algorithms), which represent distinct processes within the digital information ecosystem; Several proposed mechanisms summarized in this table remain conceptual and have not yet been directly evaluated in vaccine-specific empirical studies.

Table 2
Public health threats of artificial intelligence (AI)-driven infodemics and potential responses
Threats Potential public health consequence Potential response (primary implementing actors)
AI-generated vaccine misinformation Potential decline in vaccine confidence and increase in vaccine avoidance Strengthen transparent risk communication and trust-building through public health authorities, healthcare professionals, and community leaders
Deepfake health communication Greater perceived authenticity of false health information and erosion of trust in public health institutions Improve public capacity to recognize manipulated content through digital health literacy programs implemented by educational institutions, public health agencies, and community organizations
Algorithmic amplification and automated dissemination Rapid, widespread dissemination of misleading health information Collaborate with technology platforms to improve algorithmic transparency, content monitoring, and cooperation with independent fact-checking organizations
AI-driven infodemics during infectious disease outbreaks Delayed adoption of protective behaviors, reduced vaccination uptake, and increased outbreak vulnerability Develop real-time infodemic surveillance systems that integrate digital epidemiology with risk communication through collaboration among public health authorities, technology platforms, academic institutions, and healthcare organizations
Cross-border dissemination of AI-enabled misinformation International spread of misinformation that may undermine coordinated pandemic responses Strengthen international collaboration through WHO infodemic management, EPI-WIN, and cooperation among WHO, national public health agencies, and international public health organizations
Ethical and governance challenges in AI surveillance Potential conflicts between misinformation control and individual rights Establish governance frameworks that balance public health protection, privacy, transparency, accountability, and freedom of expression, with leadership from governments, regulatory authorities, public health agencies, and technology platforms

The proposed public health responses represent the authors’ synthesis of the literature reviewed in this Perspective and are aligned, where appropriate, with established World Health Organization (WHO) approaches to infodemic management, including the WHO infodemic management framework and related initiatives such as Information Network for Epidemics (EPI-WIN); Several proposed strategies have not yet been directly evaluated as interventions targeting AI-generated vaccine misinformation.

Figure & Data

References

    Citations

    Citations to this article as recorded by  

      When Deepfakes Meet Herd Immunity: Artificial Intelligence, Vaccine Avoidance, and the Next Infodemic Threat
      When Deepfakes Meet Herd Immunity: Artificial Intelligence, Vaccine Avoidance, and the Next Infodemic Threat
      Characteristics Traditional vaccine misinformation AI-enabled vaccine misinformation
      Content creation Created manually by individuals or organized groups Synthetic text, conversational chatbot content, synthetic audio, and deepfake images or videos can be generated at scale with minimal human effort
      Production cost and speed Constrained by human resources, technical expertise, and production time Produced rapidly with substantially lower technical and resource requirements
      Dissemination Spread primarily through interpersonal communication and conventional social media sharing Dissemination may be amplified by coordinated bot networks and platform recommendation algorithms, which are distinct from AI-based content generation
      Audience targeting Broad messaging with limited personalization AI systems can generate narratives tailored to cultural, political, religious, or health-related contexts
      Perceived authenticity Depends largely on conventional text or multimedia formats May be increased by synthetic text, audio, images, and videos that appear authentic
      Potential public health implication Localized effects on vaccine confidence and uptake Possible reduction in effective population immunity if AI-enabled misinformation contributes to widespread vaccine avoidance
      Threats Potential public health consequence Potential response (primary implementing actors)
      AI-generated vaccine misinformation Potential decline in vaccine confidence and increase in vaccine avoidance Strengthen transparent risk communication and trust-building through public health authorities, healthcare professionals, and community leaders
      Deepfake health communication Greater perceived authenticity of false health information and erosion of trust in public health institutions Improve public capacity to recognize manipulated content through digital health literacy programs implemented by educational institutions, public health agencies, and community organizations
      Algorithmic amplification and automated dissemination Rapid, widespread dissemination of misleading health information Collaborate with technology platforms to improve algorithmic transparency, content monitoring, and cooperation with independent fact-checking organizations
      AI-driven infodemics during infectious disease outbreaks Delayed adoption of protective behaviors, reduced vaccination uptake, and increased outbreak vulnerability Develop real-time infodemic surveillance systems that integrate digital epidemiology with risk communication through collaboration among public health authorities, technology platforms, academic institutions, and healthcare organizations
      Cross-border dissemination of AI-enabled misinformation International spread of misinformation that may undermine coordinated pandemic responses Strengthen international collaboration through WHO infodemic management, EPI-WIN, and cooperation among WHO, national public health agencies, and international public health organizations
      Ethical and governance challenges in AI surveillance Potential conflicts between misinformation control and individual rights Establish governance frameworks that balance public health protection, privacy, transparency, accountability, and freedom of expression, with leadership from governments, regulatory authorities, public health agencies, and technology platforms
      Table 1 Comparison of traditional and artificial intelligence (AI)-enabled vaccine misinformation across content generation, dissemination, and potential public health implications

      AI-enabled misinformation includes both generative AI technologies (e.g., synthetic text, conversational chatbots, synthetic audio, and deepfake images or videos) and automated dissemination mechanisms (e.g., coordinated bot networks and recommendation algorithms), which represent distinct processes within the digital information ecosystem; Several proposed mechanisms summarized in this table remain conceptual and have not yet been directly evaluated in vaccine-specific empirical studies.

      Table 2 Public health threats of artificial intelligence (AI)-driven infodemics and potential responses

      The proposed public health responses represent the authors’ synthesis of the literature reviewed in this Perspective and are aligned, where appropriate, with established World Health Organization (WHO) approaches to infodemic management, including the WHO infodemic management framework and related initiatives such as Information Network for Epidemics (EPI-WIN); Several proposed strategies have not yet been directly evaluated as interventions targeting AI-generated vaccine misinformation.


      JPMPH : Journal of Preventive Medicine and Public Health
      TOP