Figure 4

(a) ICEMS’s composite-score in realistic task. The vertical bars represent standard errors. There was no significant difference between three feedback groups. (b) Cognitive load. Groups are color-coded (see the legend). The vertical bars represent standard errors. Participants who received real-time AI instruction reported significantly higher extraneous load than those received in-person expert instruction. There were no significant differences between groups concerning intrinsic load and germane load. (c) Blinded expert OSATS rating. Horizontal lines represent statistically significant differences (p < .05). Vertical bars represent standard error.