2026
By Jinyi Yang
Conversational AI is becoming a normal part of higher education, but its placement in the learning process is rarely treated as a design choice. This project asks whether Socratic AI should meet students before or after formal instruction. In an online experiment, participants complete the same Communication Science research-design lesson and AI-supported problem-solving task in different orders: problem solving before instruction or instruction before problem solving. Drawing on Sinha and Kapur (2021) and Klepsch et al. (2017), the experiment examines whether sequence shapes extraneous and germane cognitive load and research-design transfer. For Communication Science, the project brings AI tutoring into message design by asking how the timing of dialogue shapes learning.