Abstract
Introduction. The integration of artificial intelligence (AI) into everyday human interactions, particularly through chatbots, has significantly increased in recent years. AI-based chatbots successfully mimic human conversations by understanding user requests and responding appropriately. Improving the effectiveness of educational and professional activities related to text processing in these conditions seems extremely relevant and predetermines the goal of our study: comparing the effectiveness of interaction with ChatGPT when formulating queries in various communication strategies. Procedure and methods. The study on the effectiveness of interacting with ChatGPT was conducted using a natural experiment scheme with preliminary training. The subjects of the study were 34 employees (15 men, 19 women, aged 25-36; M = 31.6; SD = 1.8) from the marketing department of the company "Technologies of Trust". These employees performed rewriting tasks by interacting with ChatGPT within the framework of specially adapted communication strategies (persuasive and cooperative) and evaluated the chatbot's usability and the quality of the text generated by ChatGPT. The Wilcoxon signed-rank test for dependent samples was used to identify the effects of communication strategies on interaction efficiency, and exploratory factor analysis with subsequent varimax rotation was employed to identify latent variables of interaction under different communication strategies. Research Results. Comparison of the results of the ChatGPT interaction performance evaluation when using 2 strategies (cooperative and persuasive) showed that for all 11 usability indicators, there is a significant advantage of the cooperative communicative strategy at the significance level of p<0.05. However, on 6 indicators of text quality obtained through rewriting, the persuasive strategy shows advantages and on 4 of them at a significance level of p<0.05. In the course of the study it was found that the factor solutions for the studied communicative strategies differ significantly: the key factor in using the persuasive strategy is speed and accuracy in achieving goals, while in the cooperative strategy the adequate response of chatbots to objections is critical. Conclusion. The results obtained in the study contribute significantly to the understanding of communication with AI chatbots, highlighting the nuances of interacting through linguistic means typical for human communication. The findings demonstrate the complexity of choosing the optimal strategy for interacting with ChatGPT and uncover the factors that determine the features of subjective evaluation of the interaction.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call