The essay examines how transaction-generated information—data produced through people’s online activities—can feed automated systems that sort people into categories. Those assessments may influence access to finance, housing, health care, education, social services and other opportunities, not only targeted advertising.

It argues that errors and built-in biases can enter at multiple stages, from data collection and measurement to prediction and decision-making. The essay compares data-protection and privacy rules with individual or class lawsuits under tort law, as well as regulatory models drawn from consumer product safety and environmental protection.