Under these conditions, the consistency that could make algorithmic management fairer instead entrenches the firm’s advantage. Machine learning systems can monitor worker performance, tracking outputs, time allocation, and patterns as workers complete tasks extremely closely. Firms have always had advantages over workers in terms of information, but AI greatly reduces the cost of acting on those advantages at scale.52 This dynamic is most apparent in platform companies in which algorithms can make automated decisions about task assignment, pricing, and account status with limited human oversight. They can also deactivate workers’ accounts, with limited options for appeal. Dynamic pricing that modifies prices constantly depending on supply and demand and other conditions shifts income variability from the platform onto the worker. Platform communication structures are often one-sided: Firms can reach individual workers directly, but workers cannot easily communicate with one another through the platform, making it harder to organize collectively. Box 6.2 highlights how these mechanisms play out in ride-hailing platforms and how courts and regulators have begun to respond in high-­ income countries. In developing countries, most existing employment is in the informal sector. In Sub-Saharan Africa, informal employment accounts for 85 ­percent of total employment.53 When a­lgorithmic management enters informal sectors in these countries through apps for taxi drivers, delivery workers, or domestic workers, among others, it formalizes some aspects of work without bringing the protections traditionally associated with formalization, such as employment contracts, social security, and grievance mechanisms. Box 6.2 How AI amplifies the market power of ride-hailing and delivery platforms Ride-hailing platforms have become the most litigated case of algorithmic labor management. Uber uses artificial intelligence (AI) algorithms to manage approximately 9 million drivers worldwide, controlling every aspect of the job, including who gets which rides, how much the fare is, and whether a driver’s account remains active. The platform modifies prices ­constantly by using machine learning algorithms to forecast demand. Drivers have no information about how these decisions are made, how the drivers themselves are rated, or why they are offered particular rides at specific prices. Each driver only sees their own transactions, but the platform tracks every transaction for every driver. This information advantage translates into leverage for the platform. Drivers have no control over the algorithms. Drivers must maintain scores above a certain threshold to avoid deactivation under Uber’s rating system. The platform continuously tracks acceptance and cancelation rates, response times, driving habits, and location. Algorithms adjust fares in real time. Performance categories determine which drivers get priority access to trips, creating what a Dutch court described as “a financial incentive and a disciplining and instructing effect.” a (Box continues next page) 214 World Development Report 2026