Cognitive Offloading of an Attentionally Demanding Task to a Social Robot
Basil Wahn, Khadija Bawari, Eva Wiese · 2025
Offloading tasks to artificial agents, such as ChatGPT, has become commonplace, yet little research examined offloading of attentionally demanding tasks. Prior work investigating how humans offload attentional demands to algorithms and humans found that people prefer an equal split of attentional demands when interacting with other humans but not when interacting with algorithms. The question arises how offloading decisions are made when interacting with embodied AI in form of a social robot with human-like characteristics: Would they treat it in the same way as a human or like an algorithm? To investigate this question, participants performed an attentionally demanding task (i.e., the multiple object tracking task, MOT), which they could (partially or fully) offload to the social robot MAKI, whose task accuracy was either known (Experiment 1) or unknown (Experiment 2) to the participants. In both experiments, participants offloaded a significant number of the tracking load to the robot, which improved the participants’ own tracking accuracy; knowing the robot’s task accuracy, on the other hand, did not impact offloading behavior or performance gains. Interestingly, participants did not split attentional demands equally with the robot –like they do with human interaction partners– aligning with results from previous studies when humans offloaded attentional demands to algorithms. Questionnaire data indicates that this might be due participants perceiving the robot as rather machine-like. In conclusion, our results indicate that when it comes to cognitive offloading behavior, a social robot is treated similarly to an algorithm rather than a human interaction partner.