Poll links frequent AI health queries to higher belief in vaccine myths

Adults in the U.S. who consult artificial intelligence chatbots for health advice at least once a week are more likely to accept several vaccine-related claims that health authorities have rejected, according to a poll released Tuesday by health research firm KFF.

The survey, conducted in May among a representative sample of 2,480 adults, found that 35 percent of weekly AI users said it was "definitely or probably true" that measles, mumps, and rubella vaccines cause autism in children. The corresponding figure was 20 percent among adults who never use AI for health information.

Survey details Twenty-nine percent of frequent AI users said mRNA vaccines can change a person's DNA, compared with 20 percent of non-users. Twenty-two percent of frequent users said the measles vaccine is more dangerous than the measles virus itself, versus 15 percent of those who do not consult AI.

The same poll found that adults who seek health information on social media at least weekly were more than twice as likely as non-users to accept the autism link, at 37 percent versus 16 percent.

Demographic patterns Lower-income respondents and those without college degrees were more likely to turn to social media for health advice. Respondents in households earning above $90,000 or who hold college degrees were more likely to use AI tools.

The poll did not identify which AI models respondents used. Different systems can return different answers depending on training data and company policies for handling medical queries.