Voices from the algorithm: Large language models in social research

Abstract

Research on energy and society often relies on online data collection. In particular, there has been an increase in the use of online techniques such as video software for qualitative research since the pandemic. We suggest that the rapid growth in generative AI and Large Language Models such as Chat-GPT mean that they may be utilised by research participants; particularly in research where participants may be less knowledgeable about the topic under discussion, such as emerging energy technologies. Drawing on examples from recent research, we argue that social scientists need to be cautious in assuming that the voices of our participants are genuinely theirs.