The ‘Trolley Problem’, a thought experiment in Philosophy, became characteristic of the ‘ethics of autonomous driving’, a short-lived but powerful narrative that emerged alongside the hype about driverless cars. The Problem remains a handy classroom exercise in drawing out the differences between utilitarian and deontological approaches to ethics in Analytic Philosophy. However, the thought experiment also determined the imagination of what AI-infused artifacts should be capable of, and suggested that ethical decision-making become a data-driven enterprise rather than a human, social, or individualised practice. Auto-Correct tracks the language, materiality, and culture of epistemic tools like the Trolley Problem and Moral Machine that establish safety and automobility as problems to be solved by the driverless car. This book blends critical studies of technology, culture, and society to analyse how the driverless car generates new forms of algorithmic and data-driven governmentality, crashes, victims, values, and accountability mechanisms. Auto-correct examines recent crashes and the ‘ethics of autonomous driving’ discourse through the lens of three co-constitutive cultural ontologies of the driverless car: as an AI/robot imaginary, a big data infrastructure, and a conventional 20th century automobile. In this way, this book argues for attention to ethics and values not just as outputs of AI systems but as what constitutes our social and technological presents and futures.