STROLLING CITIES. Visual Poetry generated by Artificial Intelligence

Strolling cities is a project developed in collaboration with MIT-IBM Watson AI Lab, MindEarth, and Material Balance research group. It is presented as a video installation at the 17th International Architecture Biennale di Venezia.

Strolling Cities unveils the naked, materially seductive form of 9 Italian cities – Milan, Como, Bergamo, Venice, Bologna, Genoa, Rome, Catania, Palermo – by means of millions of photos taken by the students of Politecnico di Milano, AUIC School as a part of the workshop during the recent lockdowns (’20/’21) that show the urban space as an unfiltered landscape of walls, streets, and buildings. Returned to the immanence of their materiality, cities abandon their stereotyped semantic contents to embrace a new dimension of extreme elusiveness. A generative A.I. model trained with these images creates perpetually moving video-paintings, whose indefinite contours suggest a potential transformation of urban places, once ascribed to specific social functions, into open spaces available to countless (re)writings. The observer strolls while standing still, and the city changes in front, generating a unique cognitive experience, questioning and re-imagining the space at once.

Mauro Martino and Luca Stornaiuolo created the A.I. model used in generating the images and Voice-to-City technology. Thanks to Voice-to-City, it is possible to manipulate the urban landscape with the voice, using any vocal input, from simple utterances to more complex phrases. To implement the experiment, then, we decided to use words arranged in the most synthetic and perceptually dense literary form, rooted in the sphere of aurality (i.e. in both orality and literacy, auditory and visual sensations): poetry. When fed to the A.I. model, the chosen texts foster unpredictable reactions, generating an estrangement effect that broadens immensely the common imaginary of cities. Because the model is trained to recognize patterns – that is, iterations instead of singular, unique items – the urban images conveyed by the poetic words are associated with a series of common, unremarkable places in the cities, while their more obvious landmarks are obliterated.

Each one of the 9 Strolling Cities appears perfectly recognizable – and plunged into an intimate atmosphere of belonging – to whoever knows them: in the repeated patterns of the urban space (familiar streets, everyday squares, shop windows, walls, doors, etc.) the system finds un unconventional but unmistakable beauty. To enhance this productive aspect of estrangement, we interspersed the 9 urban sections of the project with excerpts from an extraordinary manual of urban complexity, dissonance, and resilience: Le città invisibili (Invisible Cities) by Italo Calvino.

Learn more and play with the interactive module at the

Mauro Martino – MIT IBM Lab, AI Artist
Ingrid Paoletti – PoliMI, Dip. ABC, Material Balance Research Coordinator
Maria Anishchenko – PoliMI, Dip. ABC, Material Balance Research
Seyma Adali – PoliMI, Dip. ABC, Material Balance Research
Luca Stornaiuolo – Generative Models
Dalila Colucci – Selection of Poetry
Emanuele Strano – MindEarth

Narrated by
Federica Fracassi
Michele Di Mauro

Journey Team
Begum Sardan, Esra Kagitci, Giulia Annese, Giulia Grassi, Gokhan Dede, Guijia Zhao, Guitsa Herro, Huihong Zhai, Ilaria Donadel, Luca Spolaore, Manuel Zangirolami, Martina Ghidini, Michelle Rodriguez, Olga Beatrice Carcassi, Paoloandrea Paganotto, Sebastiano Nespoli, Shangyu Lou, Yinglei Li.

In collaboration with
Politecnico di Milano – Material Balance Research
MIT – IBM Watson AI Lab