<p>Artificial intelligence (AI) is transforming survey research, offering powerful tools like large language models (LLMs) to analyze human beliefs, opinions, and behaviors. As researchers increasingly adopt these technologies, they face several methodological and ethical challenges. This article explores how LLMs can enhance key stages of the survey research process within communication studies while critically assessing their limitations. Through examples related to questionnaire development and open-ended response analysis, we illustrate both the promise and pitfalls of AI-assisted surveys. Furthermore, we underscore risks regarding synthetic data generation that demand further scrutiny. Thoughtful adoption, rather than uncritical enthusiasm, is essential to ensuring that LLMs strengthen, rather than undermine, the rigor and reliability of survey-based research in communication.</p>