COVID-19 Vaccine Acceptance in the US and UK in the Early Phase of the Pandemic: AI-Generated Vaccines Hesitancy for Minors, and the Role of Governments

Gabriel Lima, Meeyoung Cha, Chiyoung Cha, Hyeyoung Hwang

Abstract

This study presents survey results of the public's willingness to get vaccinated against COVID-19 during an early phase of the pandemic and examines factors that could influence vaccine acceptance based on a between-subjects design. A representative quota sample of 572 adults in the US and UK participated in an online survey. First, the participants' medical use tendencies and initial vaccine acceptance were assessed; then, short vignettes were provided to evaluate their changes in attitude towards COVID-19 vaccines. For data analysis, ANOVA and post hoc pairwise comparisons were used. The participants were more reluctant to vaccinate their children than themselves and the elderly. The use of artificial intelligence (AI) in vaccine development did not influence vaccine acceptance. Vignettes that explicitly stated the high effectiveness of vaccines led to an increase in vaccine acceptance. Our study suggests public policies emphasizing the vaccine effectiveness against the virus could lead to higher vaccination rates. We also discuss the public's expectations of governments concerning vaccine safety and present a series of implications based on our findings.

Introduction

As the global community raced to develop COVID-19 vaccines in early 2020, researchers and policymakers faced a dual challenge: the technical hurdle of rapid development and the social hurdle of vaccine hesitancy. While vaccines are the primary tool for achieving herd immunity, public trust is fragile. This study investigates how the public perceives COVID-19 vaccines when developed using artificial intelligence (AI) versus traditional human methods. While AI involvement surprisingly didn't trigger additional hesitancy, the research uncovered a significant "protection gap": parents were far more reluctant to vaccinate their children than themselves or the elderly, and they held governments primarily liable for any potential side effects.

The authors specifically set out to explore three variables that might influence vaccine acceptance:

  1. The "Who": Does it matter if a vaccine is developed by human researchers, an AI system, or a collaboration between the two?
  2. The "How": How does the framing of news (emphasizing high effectiveness vs. reporting mild side effects) shift public opinion?
  3. The "Whom": Are people equally willing to vaccinate themselves, their children, and the elderly?

Methods and Data

The researchers conducted a cross-sectional study using a between-subjects experimental design, recruiting a representative quota sample of 572 adults from the US and UK via the Prolific crowdsourcing platform. The sample was balanced for sex, age, and ethnicity to reflect the demographic distributions of both countries.

Demographic attributes of the study participants in the US and UK.

The core of the experiment involved "news-like vignettes" (simulated news articles). Participants were randomly assigned to read different scenarios:

  • Development Source: Human-developed, AI-developed (using a hypothetical program called SAM), or a Human-AI collaboration.
  • Framing: A "positive" vignette (emphasizing high effectiveness), a "negative" vignette (mentioning mild side effects like headaches), or a "control" (no mention of results).

The team used ANOVA and post hoc pairwise comparisons to analyze changes in vaccine willingness and used the Medical Maximizer-Minimizer Scale (MMMS) to categorize participants' general predispositions toward medical intervention.

Key Results and Insights

The study captured data in April 2020, providing a baseline of public sentiment before mass vaccination campaigns began. One critical baseline measure was the level of concern about COVID-19 among the participants, which ranged from little to much concern.

Cumulative Distribution Function (CDF) of concern about COVID-19 among participants.

The AI "Non-Factor"

Contrary to fears that "black box" algorithms might spook the public, the involvement of AI in vaccine discovery did not significantly alter acceptance rates. People seemed to view AI as just another tool in the lab, provided humans were still involved in the final synthesis.

The Children Gap

This was the most striking finding. Participants were significantly more hesitant to vaccinate children than themselves or the elderly. Willingness to vaccinate children (95% CI [-0.083, 0.036]) was significantly lower than for self ([0.148, 0.268]) or elders ([0.150, 0.270]). This suggests that "protective" instincts for minors may outweigh the perceived benefits of herd immunity in the public's mind.

Information Framing

Interestingly, both positive and negative framing (mentioning mild side effects) increased vaccine willingness compared to the control group that received no information. Positive framing increased willingness by a contrast of 0.162 (p < .001) over the control, while negative framing increased it by 0.100 (p < .001). This suggests that transparency about even minor risks can build more trust than a total lack of information.

Responsibility and Agency

When considering potential side effects, the public attributes responsibility to the AI system and its programmers similarly. However, the government is held to a much higher standard of liability and punishment than the researchers or the AI companies.

Discussion and Limitations

The study's primary limitation is its timing. Conducted in April 2020, it reflects a period of high uncertainty and "hypothetical" vaccines. Public sentiment shifted significantly once real-world data and mandates emerged. Additionally, the study relied on self-reported willingness rather than actual behavior.

Future work should investigate why parents are specifically more reluctant regarding their children (e.g., fear of long-term developmental effects) and explore how AI-related misinformation might interact with the "SAM" style of AI drug discovery. Testing these variables in different cultural contexts outside the US and UK is also essential for a global understanding of vaccine acceptance.

Summary updated

Abstract

This study presents survey results of the public's willingness to get vaccinated against COVID-19 during an early phase of the pandemic and examines factors that could influence vaccine acceptance based on a between-subjects design. A representative quota sample of 572 adults in the US and UK participated in an online survey. First, the participants' medical use tendencies and initial vaccine acceptance were assessed; then, short vignettes were provided to evaluate their changes in attitude towards COVID-19 vaccines. For data analysis, ANOVA and post hoc pairwise comparisons were used. The participants were more reluctant to vaccinate their children than themselves and the elderly. The use of artificial intelligence (AI) in vaccine development did not influence vaccine acceptance. Vignettes that explicitly stated the high effectiveness of vaccines led to an increase in vaccine acceptance. Our study suggests public policies emphasizing the vaccine effectiveness against the virus could lead to higher vaccination rates. We also discuss the public's expectations of governments concerning vaccine safety and present a series of implications based on our findings.
BibTeX
@misc{lima2020covid,
  author = {Gabriel Lima and Meeyoung Cha and Chiyoung Cha and Hyeyoung Hwang},
  title = {{COVID-19 Vaccine Acceptance in the US and UK in the Early Phase of the Pandemic: AI-Generated Vaccines Hesitancy for Minors, and the Role of Governments}},
  year = {2020},
  eprint = {2006.08164},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2006.08164},
}
Figure