A new survey from KFF (Kaiser Family Foundation) shows a strong association between using artificial intelligence (AI) tools for health purposes and acceptance of vaccine‑related falsehoods. The study was conducted in the United States on a representative sample of 2,480 adults.
The research found that adults who use AI at least once a week for health information are more likely to accept certain false claims. For example, 35 percent of adults who use AI at least weekly said they are certain or think it likely that the MMR vaccine (measles, mumps, rubella) causes autism in children; among those who never use AI for health purposes, that share is 20 percent.
Similar patterns appear for other common myths:
- 29 percent of frequent technology users believe that mRNA‑based vaccines can alter human DNA, compared with 20 percent among non‑users.
- The claim that the measles vaccine is dangerous is accepted by 22 percent of regular AI users versus 15 percent of those who avoid chatbots.
Researchers adjusted for demographic and other factors such as age, race, education, and political affiliation; the association between AI use and acceptance of these myths remained clear after these controls.
Using social media for health information is also strongly linked to vaccine skepticism: among people who visit social platforms at least weekly for health content, 37 percent accept the MMR‑vaccine myth versus 16 percent among non‑users.
User groups differ markedly, however: lower‑income and less‑educated populations tend to rely more on social media, while higher‑income, college‑educated households are more likely to turn to AI tools. The study did not specify which AI models respondents used.
That omission matters because different chatbots and AI systems vary in their propensity to produce misinformation, depending heavily on the datasets used for training and on how systems handle scientifically contested or divisive questions. KFF also noted—building on an earlier March survey—that roughly one‑third of adults already seek health advice from AI; OpenAI has acknowledged that health is one of ChatGPT’s most common use cases and that the system receives hundreds of millions of such queries weekly.
The survey highlights that information sources and modes of technology use can influence public understanding and trust around vaccines. At the same time, the lack of detail about specific AI models indicates a need for further, targeted research to determine which systems or behaviors most contribute to the spread of misinformation and how to mitigate that risk.
As background, the article recalls that the false link between the MMR vaccine and autism dates to a study published in The Lancet in the 1990s that was later fully retracted.
Conclusions
KFF’s findings suggest both AI tools and social media are associated with greater acceptance of vaccine myths, though they reach different demographic groups. Because the survey did not identify the particular AI models used, additional studies are required to pinpoint how individual systems or practices drive misinformation and what interventions might reduce its impact.
