The much feared (or revered) automation of management is really the decomposition of a unified job into a variety of tasks – some done by machines, others by workers. A common criticism of Uber, for instance, is that the company has entirely replaced managers with algorithms. In reality, much of the managerial role is now split between algorithms and a crowd of workers on platforms like Appen. Consequently, management starts to resemble something entirely unfamiliar. The manager of a taxi company would usually have to supervise a team of drivers, making sure, among other things, that they are safe to be on the road and are who they say they are. Uber has famously struggled with this problem, not least because the facial recognition software it employs is prone to errors. A driver might be flagged as dangerous because their daily photo authentication fails to match their ID on record, perhaps because they shaved their beard or had a new haircut. Because the algorithm is unable to assess the driver’s credibility, Uber automatically sends a validation task to a platform like Appen. The worker who accepts the task gets thirty seconds to validate whether the driver is who they say they are. If the worker decides ‘yes’, the ride goes ahead; if ‘no’, the ride is cancelled and the driver locked out of their account. For less than a minute, the worker indirectly takes the role of Uber’s manager, in effect supervising the algorithm that supervises the labour process and makes decisions about the company’s workflow.
Phil Jones, Work Without the Worker: Labour in the Age of Platform Capitalism