Artificial intelligence is rapidly developing and being implemented in various fields, now used as an architectural design and planning tool. Online platforms, such as Dall-e, Midjourney, and Stable Diffusion, offer robots that can produce 2D drawings and 3D models corresponding to conceived space of Lefebvrian triad perceived, conceived, and lived spaces. Robots working with neural networks can think like humans, and with GAN technology, they can access architectural data faster than humans. They imagine new spaces, quickly change the scale, and develop variations. Space can be written by text since the production is based on a textual dialogue between the robot and the human. This article explores the role of human designers in establishing a dialogue with artificial intelligence, specifically with the Midjourney robot. It examines how knowledge of the history of architecture and terminology can be used as raw materials, change the quality of spatial production, and help artificial intelligence to create its own architectural style. This article proposes using semiotics as a tool in human-robot cooperation, emphasising the crucial role of architectural history and terminology to contribute to the study of AI in developing architecture.

Artificial intelligence and architecture, architectural history, theory and criticism, architectural semiotics, architectural education and technology, production of space.