Abstract

This paper reports an eye tracking study on understanding the cognitive processes of humans evaluating AI generated outputs in problem-solving tasks. We seek to understand the effect of individual differences on cognitive processing of GenAI outputs. This is because individuals differ in terms of ability to perform tasks and in their trust of Generative AI tools. We also seek to understand the difference in the cognitive processes when outputs are created by humans (experts) versus created by GenAI tools. We found that humans process GenAI outputs differently based on their level of trust on GenAI. Further, the processing of these outputs is different when they are designed by humans compared to that of GenAI. Our study highlights the importance of designing GenAI content as per individual differences and trust on GenAI tools. Our study can provide guidance on how GenAI tools should be used in the education domain.

Download

Share

COinS