Changes in attitudes toward meat consumption after chatting with a large language model

Researchers have started to explore the persuasive power of large language models (LLMs) and initial results suggest that LLMs are at least as persuasive as humans, even for controversial and moralized topics. Here, we tested whether a brief, open-ended conversation with ChatGPT-4o can change attitudes toward meat consumption, given that reducing meat consumption has become an important goal of many institutions due to its negative effects on the environment, health, and animal welfare. In a preregistered study, we found that participants (n = 101) reported a weaker commitment to eating meat and weaker beliefs that eating meat is necessary and natural, but not weaker beliefs that eating meat is normal or nice, after (vs. before) their conversation with ChatGPT-4o.