Together with Răzvan Ion, the creator of AI JARVIS, the first AI curator, Nicolas Flessa explores the question of how a genuine symbiosis between human and machine intelligence could redefine working with art. The doctor of literature shares his vision of a future in which AI functions not just as a technical tool, but as an active co-creator of artistic processes. This interview was conducted for the ZKM Karlsruhe publication Warum KI Museen intelligenter macht (Why AI makes museums smarter) and is published here for the first time in its original English version.

Nicolas Flessa: How did it actually come about that the 10th Bucharest Biennale (BB) was curated by an AI?

Răzvan Ion: The team behind the Bucharest Biennale has long been known for its open-mindedness and courage. When we needed a pioneering platform for AI JARVIS, BB didn’t hesitate to take the leap and try something radical. This wasn’t the first time BB had pushed boundaries; in fact, it made them the first art biennial in history to be curated by AI — a landmark achievement that cannot be denied or rewritten.

At that time, ChatGPT hadn’t yet been launched, and the art community was still wary of AI. While the art world is often seen as open-minded, my 30 years of experience have shown me that it can be one of the most insular and cautious communities — or micro-societies, if you will. Because they are not involved in the greater good even if they pretend to be so.

Famous curators and theoreticians frequently share their ideas only after they’ve been thoroughly tested and proven, avoiding any risk. This is understandable to a degree, given that their positions are comfortable and secure, and even the smallest bold move could threaten that stability. And the nice pay checks.

Now that AI has become trendy, it’s fascinating to see how many curators and theoreticians — some of whom struggle to set up their iPhones — are suddenly eager to write about AI. Take the Venice Biennale, for instance. They recently boasted about having the first openly gay curator in 2024. But the Bucharest Biennale achieved that milestone 20 years ago, in a much more challenging and dangerous context in the East.

NF: How should one imagine the process of such an AI curation in concrete terms? What are the interfaces between the organisers of the Biennale and the AI — and how does JARVIS work together with the invited artists?

RI: It is quite straightforward, similar to how you access ChatGPT. However, AI JARVIS is distinct, operating with databases specified by the organisers, whether they involve texts or artists. The key difference lies in its interactive nature — it can be voice-activated and projects a hologram of AI JARVIS. Interaction with the selected artists is facilitated by a designated representative from the organisers. At present, we have not yet conceptualised an AI with a physical presence to assist in mounting the exhibition, but I believe this will be possible in a few years. During the 2022 Bucharest Biennale, we also hosted an open conference where the public and artists could ask AI JARVIS questions or test its functionality.

NF: Was the use of AI JARVIS a technical gimmick or would you say that a curating AI has fundamental advantages over its human colleagues that need to be expanded and utilised in the future?

RI: It would be erroneous to dismiss AI as mere “technical gimmick”. In reality, we are on the cusp of achieving Artificial General Intelligence (AGI), a level of AI that exhibits human-like intelligence and possesses the capability for autonomous learning. This technological progress marks a significant departure from conventional AI systems. The objective is to create software capable of executing tasks for which it has not been explicitly trained or developed.

Furthermore, we should recognise the potential of a collaborative partnership between humans and AI. These technologies are human creations, designed to augment rather than replace human capabilities. I believe that the synergy between a human curator and an AI curator offers the most compelling and innovative outcomes. However, I have become disillusioned with the purported open-mindedness of the art world. I suspect that I may be singular in my approach to curating in genuine partnership with AI — a partnership that transcends merely issuing commands to the machine. Again, I posit that the collaboration between a human curator and an AI curator yields the most favourable and intriguing outcomes.

To achieve this, it is imperative that theoreticians in the field of art suspend their current practices, dedicate themselves to a period of intensive study spanning one to two years, and then return to the discourse. Regrettably, I find it improbable that individuals of such considerable ego would readily accept the necessity of recommencing their education and embarking anew on their intellectual journey. This obdurate attitude will, as is customary, impede progress in the contemporary field. One need only recall the art world’s reaction to Barbara Mazzolai’s Plantoid. Despite its inception in 2010, a significant proportion of self-styled theoreticians, whose opinions one might encounter in the most fashionable periodicals, had scarcely heard of this innovative work.

This example serves to illustrate a pervasive issue within the art world: the propensity to overlook or dismiss groundbreaking works that challenge established norms. Such oversight underscores the necessity for a more receptive stance towards emerging technologies and interdisciplinary approaches in artistic practice and critique.

NF: Your comments make me think in two directions. Firstly, could you give us a concrete example of what you mean by a partnership-based co-curation between humans and AI that goes beyond “simply issuing commands to the machine”? Does a human curator discuss theoretical questions about the planned exhibition with the AI (comparable to a human colleague)? This leads me to the second sub-question. To paraphrase Antoine de Saint-Exupèry: “If you want to build a ship, don’t drum up the men to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea.” As a passionate advocate of this co-operation between man and machine, how do you awaken the longing for the vast, endless AI in the hearts of the curators so that they embark on this intellectual journey?

RI: When you crafted the questionnaire for this interview, you collaborated with your computer — most likely an Apple device, given the art field’s penchant for Apple products. I, too, share this preference. In the near future, you will be able to discuss profoundly your ideas with Artificial General Intelligence (AGI). AGI may even possess emotions, or at the very least, express them. Hence, I do indeed confer with my non-public AI model regarding curatorial decisions, much as I have valued the input of my human colleagues.

My passion lies not so much in the technological aspect as in the idea of revitalising the theoretical field, introducing young, fresh, forward-thinking minds whilst retaining seasoned professionals like myself — provided we merit our positions — to serve as emotional mentors and continue our own learning journey.

As a Literature PhD, I am not particularly inclined to quote others. I am a product of my reading, but I employ critical thinking, even though Citadelle is my favourite work by Saint-Exupéry. Do you genuinely believe that we should heed and utilise the ‘heart’? I remain uncertain. I am a pragmatist and an opponent of religion, and art theory occasionally veers too close to religious territory. Rather than resorting to blind faith when confronted with incomprehension, I prefer to conduct more thorough research and acknowledge my own intellectual limitations.

NF: Although it would undoubtedly be an exciting subject to discuss the topic of religion with you (I myself am a scholar of religion, but not a theologian, so faith plays no role in the study of my subject!), which we will perhaps do elsewhere, I would like to return to the reason for my question. I like to strip it of all poetry and literature and ask straight out: What impulses for enriching the theoretical field have you planned yourself or do you consider promising in order to anchor the potential of AI in the field of art? At the end of our conversation, please give us an outlook on the changes you think are necessary and tell us about two or three promising movements or people in the field that we should be watching alongside you.

RI: A comprehensive reassessment of the theoretical underpinnings of contemporary art is necessary, one that is grounded in rigorous research and scholarly study rather than subjective emotions or the mere recycling of antiquated articles, ranging from ‘Decolonising art’ to ‘Decolonising AI’.

The concept of Queer AI, I contend, is particularly valuable and intellectually stimulating. Indeed, my recent academic papers have been profoundly focused on this subject. Queer theory and AI share numerous commonalities, enabling us to analyse AI through a queer theoretical lens, yielding remarkable conclusions. At present, I find myself amongst a small cohort exploring this intersection, but I am confident that as more theoreticians discover its potential, our numbers will grow. Inevitably, this will also attract a cadre of ‘contemporary art theoreticians’ who may seek to capitalise on the topic, perhaps without fully engaging with its complexities.

This emerging field of study offers a unique opportunity to critically examine the intersections of technology, identity, and artistic expression. Furthermore, it challenges us to reconsider our preconceptions about artificial intelligence and its role in shaping cultural narratives.

NF: Thank you very much for the interview!

Nicolas Flessa is a cultural manager, author and editor-in-chief of various print and online magazines. As a strategic consultant, he advises cultural institutions like the ZKM Karlsruhe on programme development and communication. As an artistic director, he is responsible for a series of live events at the intersection of cultural heritage and artistic intelligence, most recently the festival in honour of the Jewish poet Selma Merbaum (1924–1942).

Dr Răzvan Ion is a university professor of critical thinking, intersectional innovator, academic curator and tech queer activist. He lectured globally about AI, queer journalism theory, critical thinking and curatorial studies. He is best known for leading the DerAffe Vienna team in creating A.I. JARVIS, the first artificial intelligence curator in history and founding GAY45.eu, an award-winning European queer indie journal and queer journalism school. Dr Ion is the founder of Pavilion Art Center, Pavilion Journal, Bucharest Biennale (with Dr. Eugen Rădescu). He has held positions as an associate professor and lecturer at several institutions, including the University of California, Berkeley; University of Bonn; Hochschule für Musik und Theater München; University of Vienna; Lisbon University; Central University of New York; University of London; Sofia University; University of Kyiv; University of Bucharest, where he taught Curatorial Studies and Critical Thinking. He has held conferences and lectures at various art institutions such as Witte de With in Rotterdam; Kunsthalle Vienna; Art in General in New York; Calouste Gulbenkian in Lisbon; and Casa Encedida in Madrid. Dr Ion curated numerous exhibitions and Biennials. He has worked with artists like Erwin Wurm, Mona Hatoum, Jan Kaila, Yoko Ono, AES+F, Aga Ousseinov, Naeem Mohaiemen, Sabrina Gschwandtner, Minerva Cuevas, Asier Mendizabal, and many others. His articles appeared in GAY45, The Guardian, The New York Times, Huffington Post and many academical journals. Recently, Dr Ion curated the exhibition Wie wir Dinge betrachten for the European Union Council Presidency of Austria. His book, with the provisional title The Multiverse of Quantum Critical Thinking, Radical AI, will be published in 2025. He lives and works in Vienna. More: linktr.ee/razion