ESRA 2027 sessions by theme

Back to Overview of Sessions

Session Details

The integration of artificial intelligence (AI) in web surveys is gaining momentum. Recent developments in large language models (LLM) enable the use of AI on various stages of the survey life cycle. Several studies have used LLMs to propose the wording of survey questions and response categories and to serve as participants in pre-test activities. Also, AI has been used for the integration of customized adaptations of the measurement instrument during data collection. Based on real-time interpretations of responses and response behavior, survey systems adapt clarification features, administer follow-up questions, and decide on a respondent-specific question wording.
New ways to integrate AI in the survey life cycle need to be evaluated in terms of data quality, respondent experience as well as data protection and integrity. Questions have to be explored regarding whether the integration of an LLM changes the overall approach to a questionnaire. Does it alter the cognitive processing of survey questions by respondents? Do these measures enhance or deteriorate the data quality and respondent experience and why? Which safeguards may be necessary to maintain high level data protection and integrity?
The integration of LLMs in surveys also raises new methodological questions. As the AAPOR report on “Responsible AI Integration in Survey Research” (2026) states, using LLMs require new validation standards and paradata for evaluation and monitoring.
The session wants to offer insights into the application of AI during questionnaire development and survey data collection and provide a forum to discuss methodological issues that surround research on AI enhanced surveys. At the same time, it aims to provide an overview of current research on the consequences of using AI during data collection in surveys for respondents, survey researchers and survey practitioners.