Odkrývání lidských hodnot: Postřehy z konverzací s AI konverzačním Agentem

Babický, Jakub

Human values are typically measured using self-report questionnaires, which may be prone to social desirability bias and might take a long time to complete. The rise of contemporary AI- powered Conversational Agents provides us with the possibility to develop a more engaging and time-efficient way to measure human values through natural conversation. While previous studies have investigated whether such agents can infer personality traits, such as the Big Five or Myers-Briggs types, little research has explored their ability to assess human values. This study examined whether Conversational Agents can identify individuals' most important values from short conversations with individuals. A total of 203 participants engaged in a brief text- based exchange with a Conversational Agent and then completed the Portrait Values Questionnaire. The values inferred by the Conversational Agent were compared with participants' self-reported values. On average, the agent identified participants' four most important values at a rate much higher than random chance. Although this study was just a first step and future research would be needed, these findings suggest that Conversational Agents may potentially offer a widely applicable, engaging, and time-efficient tool for psychological assessment of values, possibly...

guest :: login Digital Repository
		Search		Submit		Help		About