ChatGPT vs Google: Difference in AI-Enabled Iterative Generation

With novel AIGC characteristics, the emerging ChatGPT threatens Google’s online search market dominance. We plan to compare Google and ChatGPT as question answering systems (QASs), focusing on how AI-enabled iterative generation affects users’ perceived learning capability and predictability in a laboratory experiment. Specifically, we integrate three theoretical lenses: self-regulated learning theory, question answering systems (QASs), and information retrieval literature—to examine differences between Google and ChatGPT. Two new constructs are developed: AI- enabled iterative generation and perceived learning capability. Then we propose that AI-enabled iterative generation leads to better performance in all three dimensions of learning capability (i.e., elaboration, critical thinking, and self-regulation) while Google excels in perceived predictability. When QA tasks become more complex, the superiority of ChatGPT will be more prominent while Google’s strength will diminish. Overall, our expected results can provide significant implications for the new shape and direction of search services.