As AI becomes embedded in everyday work, people wonder when the results still feel like theirs. We conducted an exploratory qualitative survey asking participants to describe two recent AI‑assisted tasks—one that felt like their own and one that did not. The findings show that ownership hinges on the mode of collaboration: merely approving AI suggestions leads to disownership, while leading, iterating, or rewriting preserves ownership.
Ownership can also reside at the vision level rather than execution. Respondents reported strong ownership for tasks they could not have completed without AI, even though the machine performed the actual work. Conversely, loss of personal voice and lack of understanding of the output erode ownership.
Willingness to disclose AI use is often decoupled from pride or ownership, being shaped instead by community norms and fear of credit erasure. Based on these insights, we propose several research directions to guide AI development that supports people’s sense of authorship over their own lives.
Review