ghost work: a standalone construct

Mary L. Gray and Siddharth Suri named the construct in their 2019 book Ghost Work, defining it as the on-demand human labor (content moderation, data labeling, transcription, task work) that powers software and AI services marketed as fully automated, while the workers who perform it are kept deliberately invisible to end users and outside standard employment protections. The argument is that behind an interface that appears to run itself sits a globally distributed, piecework workforce whose contributions are hidden by design, a condition Gray and Suri call the "paradox of automation's last mile": each advance in automation creates new, unseen human tasks rather than eliminating them. Toxtli, Suri, and Savage (2021) later quantified part of the thesis on crowd-work platforms, showing that once the hidden, unpaid task overhead is counted the effective wage falls from about $3.76 to $2.83 per hour. The construct is distinctive because it names not a technology but a labor relation (the systematic concealment of human work behind claims of automation), and it sits in the sociology of work rather than in any account of the algorithms themselves.[1][2]

Unit

One instance of hidden, on-demand human labor performed behind an automated-seeming system

Recovery means

Not applicable

Measured by

Studied through platform task logs, effective-wage calculations that include unpaid overhead, and ethnographic accounts of on-demand crowd and moderation work