Italiano [English below]
Sono intervenuto alla XXVII edizione della Generative Art Conference (17–19 Dicembre 2024), che quest’anno si è svolta a Venezia presso l’UNESCO Regional Bureau for Science and Culture in Europe. Organizzata fin dalla prima edizione del 1998 grazie all’impegno straordinario di Celestino Soddu e Enrica Colabella, la conferenza ha contribuito a livello internazionale a fondare il campo di ricerca, le idee e le riflessioni sull'”arte generativa”. Un concetto che, a partire dai tardi anni ’50 del ‘900, in particolare grazie all’impiego del calcolatore e dei dispositivi informatici nel fare artistico, ha acquisito negli anni un rilievo crescente con l’evoluzione degli strumenti tecnologici e oggi con la diffusione “esplosiva” dell’Intelligenza Artificiale generativa.
L’evento, che si svolge in Italia in sedi ogni anno diverse e che comprende anche mostre e performance, per diversi relatori è divenuto un appuntamento accademico. Il sottoscritto aveva partecipato all’edizione del 2017, che si era svolta a Ravenna presso il Museo dell’Arte (MAR) e la Biblioteca Classense, di cui avevo scritto qui.
L’argomento principale di quest’anno era “Generative Art where the Beauty Voice Survives. Future Visions on Past Steps”. Come recita il sito:
Humanity & AI. The theme of this conference focuses on how to preserve the human complexity, with Generative Art and AI. Art, poetry, music, architecture, and historic cities need generative ideas for their adaptation to the deep new demands of our fast-changing time. Our field of interest is to identify these possibilities and relate the most advanced creative approaches of Generative Art and AI.
Di rilievo, e raro in eventi internazionali come questo, il fatto che gli atti siano disponibili prima dell’inizio della conferenza, sia stampati, consegnati all’atto della registrazione, sia in formato digitale, gratuitamente scaricabili online fin dalla prima edizione del 1998, costituendo un contributo molto importante per gli studi in questo ambito di ricerca (gli atti dell’edizione 2024 possono essere scaricati da qui). Un altro aspetto da sottolineare è che non ci sono keynote, né sessioni parallele che costringono a scegliere quali interventi seguire, tutti gli interventi si svolgono uno dopo l’altro nei tre giorni della conferenza.
L’arte generativa comprende oggi una grande varietà di forme espressive che vanno ben oltre l’ambito informatico, basti pensare a quelle basate sulla biologia, come per esempio le bioarti. Una varietà bene esemplificata dalle definizioni di “arte generativa” che nel corso del tempo sono state date. Per fare solo qualche esempio:
“Generative art refers to any art practice where the artist uses a system, such as a set of natural language rules, a computer program, a machine, or other procedural invention, which is set into motion with some degree of autonomy contributing to or resulting in a completed work of art.”
[Philip Galanter, “What is Generative Art? Complexity Theory as a Context for Art Theory”, International Conference on Generative Art, 2003, online, https://www.philipgalanter.com/downloads/ga2003_paper.pdf (last access 15/12/24)]
Oppure, in maniera più legata all’autorialità:
“Generative Art is the idea realized as genetic code of artificial events, as construction of dynamic complex systems able to generate endless variations. Each Generative Project is a concept-software that works producing unique and non-repeatable events, like music or 3D Objects, as possible and manifold expressions of the generating idea strongly recognizable as a vision belonging to an artist / designer / musician / architect /mathematician.”
[Celestino Soddu, Enrica Colabella, “Generative Art”, Generative Art, 2023, online, https://generativeart.com (last access 01/08/23)]
O ancora, più in generale:
“In Generative art, the artwork is generated, at least in part, by some process that is not under the artist’s direct control.”
[Margaret A. Boden, Ernest A. Edmonds, “What is generative art?”, Digital Creativity, vol. 20, n. 1-2, 2009, p.10, DOI: 10.1080/14626260902867915 (last access 15/12/24)]
Tuttavia, gli interventi alla conferenza hanno riguardato praticamente solo l’ambito informatico. Un’ulteriore osservazione è che l’impronta della conferenza e il carattere degli interventi sono stati sostanzialmente di natura tecnica-tecnologica, mentre le dimensioni storica e teorica sono apparse secondarie e trascurate. Quest’ultimo aspetto è emerso in alcuni interventi che hanno mostrato un approccio un po’ semplicistico al rapporto tra l'”oggetto artistico” e la sua dimensione storico-sociale. Segno che forse è il momento di riprendere in mano alcuni elementi importanti dell’estetica del ‘900, tra cui l’idea, più volte citata durate la conferenza, di “opera aperta” di Umberto Eco.
- Picture generated by the author of this paper with Stable Diffusion
- Referential vs non-referential images (picture by the author)
- Klaus Spiess, Emanuel Gollob, Paul Gründorfer, Jens Hauser, Chronolalia: Towards an Oscillating Truth, bio-installation, 2023
Il mio paper, fruibile negli atti, sottolinea alcuni elementi critici e problematiche dell’Intelligenza Artificiale generativa che possono rivestire interesse in ambito artistico: gli errori, il disorientamento semantico, l’imprevedibilità, l’incoerenza e gli aspetti inerenti al copyright. Poi si sofferma sulla dimensione di simulazione degli algoritmi, e in particolare sul fotorealismo che è possibile ottenere dalle immagini di IA generativa, sulla loro “non referenzialità” opposta alla “referenzialità” delle immagini fotografiche, riprendendo un concetto che avevo espresso in un mio vecchio libro (Realtà del virtuale, Bologna, CLUEB, 1993, seconda edizione 2015). Il mio paper colloca inoltre l’Intelligenza Artificiale all’interno di un processo più generale, in atto fin dalle origini dell’umanità, di progressiva “esternalizzazione” al di fuori del corpo di funzioni e attività umane, riconoscendo che un analogo processo è in corso anche nelle arti e solleva, tra le altre, le questioni di un'”arte creata dalle macchine”, preoccupazioni etiche, problemi di originalità, di autorialità e proprietà intellettuale. Di fatto, esternalizzare la creatività a un dispositivo autonomo può ridurre e persino sostituire l’intervento umano nei processi che portano all’opera d’arte, facendo emergere nuove forme espressive. Il paper si conclude con la discussione di un’installazione artistica in cui l’intelligenza artificiale generativa non viene utilizzata per creare immagini (Klaus Spiess, Emanuel Gollob, Paul Gründorfer, Jens Hauser, Chronolalia: Towards an Oscillating Truth, 2023).
Il mio intervento alla conferenza ha presentato le attività di ricerca, le pubblicazioni e i corsi svolti dal sottoscritto nell’ambito dell’IA generativa, anche tramite Noema, con particolare riferimento all’insegnamento di questa disciplina nelle istituzioni d’arte, e ha ripreso come esempio l’installazione artistica di cui sopra in cui l’Intelligenza Artificiale generativa non viene utilizzata per creare immagini.
I attended the 27th Generative Art Conference (December 17-19, 2024), held in Venice at the UNESCO Regional Bureau for Science and Culture in Europe. Organized since 1998 through the extraordinary commitment of Celestino Soddu and Enrica Colabella, the conference has contributed internationally to establishing the research field, ideas, and reflections on “generative art.” This concept, emerging from the late 1950s, particularly through the use of computers and digital devices in artistic creation, has gained increasing prominence with the evolution of technological tools and today’s “explosive” spread of generative Artificial Intelligence.
The event, held annually in different Italian venues and including exhibitions and performances, has become an academic appointment for many speakers. I had participated in the 2017 edition, which took place in Ravenna at the Museum of Art (MAR) and the Classense Library, about which I had written here.
This year’s main theme was “Generative Art where the Beauty Voice Survives. Future Visions on Past Steps.” As stated on the event’s website:
Humanity & AI. The theme of this conference focuses on how to preserve human complexity, with Generative Art and AI. Art, poetry, music, architecture, and historic cities need generative ideas for their adaptation to the deep new demands of our fast-changing time. Our field of interest is to identify these possibilities and relate the most advanced creative approaches of Generative Art and AI.
Notably, and rare in international events like this, the proceedings are available before the conference begins, both printed and handed out at registration, and in digital format, freely downloadable online since the first edition in 1998, constituting a very important contribution to studies in this research area (the 2024 proceedings can be downloaded from here). Another noteworthy aspect is that there are no keynotes or parallel sessions forcing choices between presentations; all presentations occur consecutively over the conference’s three days.
Generative art today encompasses a wide variety of expressive forms extending well beyond the computational realm, including those based on biology, such as bioarts. This variety is well exemplified by the definitions of “generative art” given over time. To cite just a few examples:
“Generative art refers to any art practice where the artist uses a system, such as a set of natural language rules, a computer program, a machine, or other procedural invention, which is set into motion with some degree of autonomy contributing to or resulting in a completed work of art.”
[Philip Galanter, “What is Generative Art? Complexity Theory as a Context for Art Theory”, International Conference on Generative Art, 2003, online, https://www.philipgalanter.com/downloads/ga2003_paper.pdf (last access 15/12/24)]
Or, more focused on authorship:
“Generative Art is the idea realized as genetic code of artificial events, as construction of dynamic complex systems able to generate endless variations. Each Generative Project is a concept-software that works producing unique and non-repeatable events, like music or 3D Objects, as possible and manifold expressions of the generating idea strongly recognizable as a vision belonging to an artist / designer / musician / architect /mathematician.”
[Celestino Soddu, Enrica Colabella, “Generative Art”, Generative Art, 2023, online, https://generativeart.com (last access 01/08/23)]
Or more generally:
“In Generative art, the artwork is generated, at least in part, by some process that is not under the artist’s direct control.”
[Margaret A. Boden, Ernest A. Edmonds, “What is generative art?”, Digital Creativity, vol. 20, n. 1-2, 2009, p.10, DOI: 10.1080/14626260902867915 (last access 15/12/24)]
However, the conference presentations dealt almost exclusively with the computational field. A further observation is that the conference’s imprint and character of the presentations were substantially technical-technological, while historical and theoretical dimensions appeared secondary and neglected. This latter aspect emerged in some presentations that showed a somewhat simplistic approach to the relationship between the “artistic object” and its historical-social dimension. Perhaps it’s time to revisit some important elements of 20th-century aesthetics, including Umberto Eco’s frequently cited concept of “open work.”
- Picture generated by the author of this paper with Stable Diffusion
- Referential vs non-referential images (picture by the author)
- Klaus Spiess, Emanuel Gollob, Paul Gründorfer, Jens Hauser, Chronolalia: Towards an Oscillating Truth, bio-installation, 2023
My paper, available in the proceedings, highlights critical elements and issues of generative Artificial Intelligence relevant to the artistic field: errors, semantic disorientation, unpredictability, inconsistency, and copyright concerns. It then examines the simulation dimension of algorithms, particularly the photorealism achievable through generative AI images, and their “non-referentiality” versus the “referentiality” of photographic images, revisiting a concept I expressed in my old book [Realtà del virtuale (Reality of the Virtual), Bologna, CLUEB, 1993, 2nd edition 2015). The paper also places Artificial Intelligence within a broader process, ongoing since humanity’s origins, of progressively “externalizing” human functions and activities outside the body, acknowledging that a similar process is occurring in the arts, raising questions about “machine-created art,” ethical concerns, and issues of originality, authorship, and intellectual property. In fact, externalizing creativity to an autonomous device can reduce and even replace human intervention in artistic processes, giving rise to new forms of expression. The paper concludes by discussing an art installation where generative artificial intelligence is not used to create images (Klaus Spiess, Emanuel Gollob, Paul Gründorfer, Jens Hauser, Chronolalia: Towards an Oscillating Truth, 2023).
My conference presentation covered my research activities, publications, and courses in generative AI, including through Noema, with particular focus on teaching this discipline in art institutions, and revisited the aforementioned art installation where generative Artificial Intelligence is not used to create images as an example.




















