Q&A: Do AI and bogus respondents threaten polling’s future?

2026-05-12 · Source: Pew Research Center · Field: Science & Research — Social Sciences & Behavioral Studies, Research Methodology & Innovation · Depth: Novice, medium

Summary

Recent developments, including "silicon sampling" where AI generates public opinion and "bogus respondents" using AI to fake survey answers, pose significant threats to the polling industry. Courtney Kennedy, Pew Research Center's vice president of methods and innovation, clarifies that Pew does not use AI to replace human respondents due to ethical and scientific concerns, noting AI-generated opinions often stereotype groups, misrepresent Republican viewpoints, and understate public disagreement. Pew Research Center exclusively employs probability-based sampling, recruiting real people offline via mail, making it resistant to large-scale fraud seen in opt-in surveys. While opt-in polls are vulnerable to bad actors earning up to \$30,000 monthly by faking responses, Pew's method limits such gains to around \$22 monthly. Bogus respondents in opt-in surveys, who often give positive answers, have led to false conclusions, particularly concerning young adults or rare behaviors. Trustworthy polling requires rigorous, well-designed probability-based methods, including proper weighting.

Key takeaway

For research scientists and policy makers relying on public opinion data, you must critically evaluate polling methodologies. Opt-in surveys are highly susceptible to AI-generated and bogus responses, leading to skewed results, especially for specific demographics or rare behaviors. Prioritize polls employing rigorous probability-based sampling, which significantly reduces fraud risk and enhances data integrity. Always assess how respondents were recruited and if proper weighting was applied to ensure the poll's trustworthiness and avoid misinformed decisions.

Key insights

The rise of AI-generated and faked responses in polling necessitates a return to rigorous, probability-based human sampling for trustworthy public opinion data.

Principles

• Polling is fundamentally about human thought and experience.
• AI-generated opinions often stereotype and misrepresent views.
• Probability-based sampling mitigates large-scale fraud.

Method

Pew Research Center uses probability-based sampling: randomly selecting U.S. addresses, contacting via mail, and inviting a carefully selected sample to take surveys online or by phone.

In practice

• Avoid opt-in surveys for sensitive demographics or rare behaviors.
• Prioritize probability-based sampling for reliable public opinion.
• Scrutinize poll recruitment methods and weighting practices.

Topics

• Polling Methodology
• AI in Polling
• Silicon Sampling
• Bogus Respondents
• Probability Sampling
• Data Quality

Best for: AI Ethicist, Policy Maker, Research Scientist

Related on AIssential

• AI is replacing humans in responding to some surveys – but simulated opinions are not the same as public opinion — Simulating public opinion with AI models is not equivalent to measuring actual public opinion due to inherent biases and lack of real-world validation.
• Use AI This Election — AI can effectively personalize complex political research, significantly improving voter decision-making.
• Survey Statistics: more on recalled vote — Recalled vote data consistently underestimates voter volatility and is influenced by current preferences.
• Survey Statistics: it is (still) the people — Synthetic past vote (X**) improves voter support estimation by addressing missing or inaccurate recalled vote data.
• AI chatbots can effectively sway voters – in either direction — AI chatbots effectively sway political opinions by generating numerous factual claims, even if some are inaccurate.
• Adaptive auditing of AI systems with anytime-valid guarantees — Adaptive AI auditing can achieve statistical rigor using dueling hypotheses and anytime-valid e-processes, even with small, flexible datasets.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Pew Research Center.