What does it mean for a machine to be 'intelligent'? Are machines capable of being moral? Does an algorithm of perpetual peace exist? Such are the questions addressed here by philosopher Yuk Hui, who brings the work of 18th-century philosopher Immanuel Kant to bear on the artificial intelligence of today. Reading rationalism and empiricism as the Cartesian and Humean machines of Kant’s time, Hui reinterprets Kant’s philosophy as a reflection on machines and demonstrates its usefulness for finding an alternative path through contemporary debates on the nature of AI. He offers a new interpretation of the ‘epigenesis of reason’ to explore the limitations of machines that operate within the faculty of cognition but not within the faculties of feeling or desire, and explores the idea of an algorithm for perpetual peace by reinterpreting Kant’s concept of the universal through the antinomy of judgment. This concise but penetrating analytical enquiry opens up new lines of investigation for the philosophical study of artificial intelligence and resituates Kant’s critical philosophy in the epoch of generative AI.