The rise of artificial intelligence and fraudulent respondents is posing a significant threat to the accuracy of public opinion polling in the United States, according to a new Q&A from Pew Research. Courtney Kennedy, the organization's vice president of methods and innovation, addressed concerns over how these emerging challenges could undermine decades of polling methodology. She noted that the issue has grown more urgent as AI tools become increasingly sophisticated at mimicking human survey responses.
Kennedy explained that bogus respondents—sometimes called 'survey bots'—can flood online polls with fake answers, skewing results and eroding trust in data. The problem is not entirely new, but the scale has expanded with easier access to AI text generators. Pew Research has been studying these patterns to develop detection methods, though Kennedy acknowledged that defenders are often playing catch-up.
Data from Pew indicates that the prevalence of such respondents varies by survey mode, with online panels being particularly vulnerable. While exact figures were not provided, Kennedy highlighted that low-incidence rates can still distort findings on sensitive topics. The research organization is now prioritizing the refinement of quality-control checks to filter out bogus entries.
The implications extend beyond academic surveys, affecting market research, political polling, and public policy development. If left unchecked, the credibility of all survey-based insights could be compromised. Kennedy called for industry-wide collaboration to establish standards for verifying respondent authenticity.
Critics argue that fears may be overstated, pointing out that many polls already incorporate basic bot detection. They caution against overcorrecting in ways that might exclude legitimate voices from underrepresented groups.