The Digicomlab kicked off the academic year with a meeting on the usage of Large Language Models (LLMs) in responses to online surveys and experiments. Many communication science projects use online questionnaires and/or experiments: Participants can log in on their private devices and answer questions or complete tasks for scientific research. Yet, increasingly, responses are not generated by humans but by generative Artificial Intelligence relying on LLMs. Jasper Wijnen (Information Security Officer at the University of Amsterdam) discussed the risks that this poses to communication research and demonstrated potential ways of mitigating these risks.