The article demonstrates the possibility of applying such artificial intelligence systems as Large Language Models to the qualitative analysis of verbal forecasts. The purpose of the psychological study was to consider the participants forecasts in the broad context of classifications of anticipated events in relation to the individually imagined future. The experimental design included forecasting in free statements by participants (n = 149) of possible, improbable, impossible events. The sample was represented by Chinese participants who gave on-line free descriptions of such events. According to the three conditions, 447 verbal forecasts were obtained. Their analysis using the o1-preview meta-model of advanced reasoning presented in our first article revealed several significant differences in the frequencies of forecasts for the three conditions when categorizing statements predicting the future in the context of possible, impossible and unlikely events. Eleven dimensions – criteria for classifying predicted events – showed significant differences, including impossible/surreal events, realistic expectations, fantastic scenarios, creative thinking, factual thinking, everyday routine, science fiction elements, dystopian elements, supernatural events, cultural influences, and inductive reasoning. The obtained results are important for the development of hybrid research methodologies involving humans and AI and a broader understanding of artificial cognitive systems as tools for expanding human analytical capabilities.
Kornilova et al. (2025) studied this question.