The role of perceived anthropomorphism and anthropomorphic design elements in willingness to delegate writing tasks to AI: findings from a large-scale survey and a randomised controlled experiment
Teresa Luther, Maren Mayer, Joachim Kimmerle · Journal of Psychology and AI · 2026
With the growing integration of large language models (LLMs) such as ChatGPT into writing tasks, understanding factors that shape human-AI interaction and task delegation is essential. The present research investigates how user perceptions of AI systems and anthropomorphic design features shape willingness to delegate writing tasks, interaction behaviour, textual outcomes, and perceptions of AI contribution in AI-assisted writing. We conducted two complementary studies with U.S. samples: a large-scale survey (N = 1007) and a randomised controlled experiment (N = 397). The survey examined how perceptions of AI systems predict willingness to delegate writing tasks. Building on the survey, the experiment tested the causal effects of anthropomorphic design features on user perceptions, delegation behaviour, and co-created text outcomes across personal and professional writing tasks. In the survey, perceived trustworthiness and anthropomorphism predicted delegation intention, whereas social agency and perceived intelligence did not. In the experiment, anthropomorphic design increased positive perceptions, prompting frequency, adoption of AI-generated content, and analytical language use. Participants also underestimated the AI’s contribution, particularly in professional tasks and when interacting with an anthropomorphic AI. Anthropomorphic design shapes both perceptions and behavioural enactment of delegation in writing, while influencing attribution processes in human-AI collaboration.