Your next respondent might not be a human. AI agents can now read a questionnaire, answer questions convincingly, and even pause and make typos like a human. Standard checks are not enough anymore, and platforms that promise “100% human” data should be taken with a grain of salt.

What does this mean for online data collection, and what can we do?

Join us to find out how to spot synthetic respondents, which alternatives to online panels are available, and what this means for research in the social and behavioral sciences.

Introduction

The event will begin with a talk by Jasper Wijnen (FMG Research Lab), who will introduce the problem, discuss its extent, and offer solutions based on their experiences helping researchers at the Faculty of Social and Behavioral Sciences.

Researchers sharing experiences and solutions

In this session, we will have short talks by researchers who encountered AI pollution in their research. They will share how they dealt with this issue and the solutions they found.

Speakers

• Roberto Cerina (UvA, Computational social science)
• Leendert van Maanen (Utrecht University, Psychology)
• Marthe Möller (UvA, Communication science)
• Joanna Strycharz (UvA, Communication science)

Panel discussion

In this panel discussion, we will talk about the role of survey platforms in protecting data quality, discuss the risks and opportunities created by generative AI, and the future of online data collection.

Speakers

• Denny Borsboom (UvA, Psychology)
• Oriol Jover (Prolific)
• Felicia Loecherbach (UvA, Communication Science)
• Moderated by Tessel Derckx (FMG Research Lab)