Inspired by the work of Friedrich Schiller and John Hejduk, the paper proposes that an architecture education informed by the inherent aesthetic perception and cognition limits of the architect and artificial intelligence (AI) is fundamental to articulating a possible pedagogical reform in the age of GenAI. Contrary to the normative understanding of the ‘AI alignment problem’, beyond issues of ethics, it argues for a productive AI-misalignment in creativity in engendering a new AI-native aesthetics characterised by ambiguity—artificial aesthetic. Sample projects drawn from the author’s own teaching and design practice are used to illustrate artificial aesthetics at work in architectural perception, reasoning, and generation.