AI Workslop: The Moral Significance of Withholding Effort

Abstract This paper examines the moral status of “AI workslop”: AI-generated output that appears adequate but lacks the substance a task requires. Existing research has measured its prevalence and costs and suggested managerial responses, but has not addressed when producing workslop is wrong, permissible, or required. I connect AI workslop to an older problem in workplace ethics: effort-withholding, as in shirking or slacking. Workslop is often perceived as continuous with these practices because it can shift burdens onto others while appearing to spare the sender effort. But the inference from workslop to withheld effort is insecure. Workslop is individuated by output and tool use, not by intention or effort. Its distinctive profile lies in intention-independence, its uncertain relation between output and effort, and the distance AI delegation can place between workers and burdens imposed on others. I argue that responsibility for producing, preventing, and responding to workslop turns on the distribution of control. Managers shape structural conditions, including workloads, deadlines, metrics, AI policies, and training; workers control local choices about whether and how to use AI, how carefully to check its output, and how much burden they pass on. Drawing on harm-based considerations shared across moral traditions, I argue that producing workslop is wrong when it imposes substantial, avoidable, and nonredundant burdens that could reasonably have been prevented; permissible when harms are minor or structurally unavoidable; and, in narrow non-ideal cases, required. The account favors structural reform and worker support over punishment, reserving sanctions for high-stakes settings and repeated deliberate offloading.