Art and artificial intelligence in 2026: three scenarios for galleries

01First scenario: AI as an integrated tool within artistic practice

The first scenario, and probably the most likely in the short term, is a gradual integration of artificial intelligence into the contemporary artist's toolkit, alongside digital photography, video editing software and 3D printing. In this scenario, AI does not replace the artist's hand but expands the possibilities. Refik Anadol, whose immersive installations use neural networks to transform massive datasets into visual environments, illustrates this approach. His works have been shown at MoMA in New York and the Serpentine in London, and they have achieved significant prices at public sale. The artist retains intellectual and aesthetic control of the work; the machine is one instrument among many.

For the dealer, this scenario requires developing the capacity to discern among a growing number of artists who claim to use AI in their practice. The question is not whether an artist uses AI but how and why. An artist who merely generates images from text prompts without substantial intervention does not propose an artistic practice in the sense that the gallery world understands it. Conversely, an artist who programmes their own models, trains their algorithms on specific corpora, manually intervenes on generated results and situates this approach within a coherent conceptual framework deserves the dealer's attention.

Pace Gallery in New York was among the first to represent artists working with artificial intelligence, and it has distinguished practitioners whose work possesses conceptual depth from those riding a trend. Galerie Nagel Draxler in Cologne has likewise integrated digital artists into its programme while maintaining the same level of intellectual rigour as for its artists working in traditional media. These examples demonstrate that a gallery can welcome AI into its programming without compromising its reputation for curatorial rigour.

02Second scenario: the emergence of a parallel market dedicated to AI art

The second scenario envisages the formation of a distinct market, with its own galleries, its own fairs, its own collectors and its own valuation criteria. This market already exists in embryonic form: online platforms specialising in the sale of AI-generated works have appeared, exhibitions entirely devoted to algorithmic art are multiplying, and a community of collectors has formed around these practices, often drawn from the technology world rather than the traditional art world.

In this scenario, the traditional contemporary art dealer does not necessarily need to position themselves on this segment. The AI art market develops according to its own logic, with prices often lower than those of the established contemporary art market, and a clientele that values technological innovation as much as aesthetic quality. Galleries that choose not to integrate AI art into their programme are not necessarily falling behind: they are making a positioning choice consistent with their identity and the expectations of their existing clientele.

However, the dealer must remain alert to the bridges being built between these two worlds. Holly Herndon, whose work blends artificial intelligence, music and visual art, is shown in both technological contexts and prestigious contemporary art institutions. Trevor Paglen, represented by Pace Gallery, uses artificial intelligence to interrogate surveillance and facial recognition systems, situating his practice in a tradition of conceptual and political art that interests the most demanding collectors. These bridge-artists are the ones the dealer should watch closely, as they may redefine the boundaries between the two markets.

03Third scenario: a fundamental questioning of the notion of authorship

The third scenario, the most radical and the most uncertain, envisages a profound transformation of the very notion of artistic authorship. If artificial intelligence models become capable of producing works indistinguishable from those created by human artists, the question of authenticity, singularity and the value of human intervention will arise with fresh urgency. This scenario is not imminent, but it is taking shape on the horizon and the dealer who thinks long-term cannot afford to ignore it.

In this scenario, the gallery could play a more important role than ever as a guarantor of certification and authenticity. If the production of aesthetically accomplished images becomes trivial through AI, what will distinguish a work of art from a mere image will be precisely the work of contextualisation, legitimation and valorisation that the dealer performs. The artist's trajectory, the intention behind the work, the dialogue with art history, the inscription within a coherent curatorial programme: all these elements the dealer brings would become even more decisive in a world saturated with images.

Galerie Thaddaeus Ropac has always insisted on the importance of the narrative that accompanies the work, and this approach takes on a new dimension in a context where image production is potentially unlimited. Galleria Continua, with spaces spread across several continents, has built its identity on the ability to tell artistic stories that transcend the materiality of the work, and this competence will be increasingly sought by collectors confronted with an ocean of images whose provenance they can no longer evaluate.

04Practical implications for the dealer

Whichever scenario prevails, and it is probable that all three will coexist to varying degrees, the dealer must make a number of practical decisions now. The first concerns training: understanding the basics of how generative artificial intelligence works, the differences between models, the legal issues surrounding copyright and training data, is no longer optional for an art market professional. This is not about becoming an engineer but about possessing the interpretive keys necessary to evaluate the work of artists who use these tools.

The second decision concerns communication. The dealer must be able to explain to collectors why they choose, or choose not, to present works involving artificial intelligence. A reasoned and nuanced discourse reassures the collector and reinforces the dealer's credibility. The absence of discourse, by contrast, leaves the field open to received ideas and unfounded fears that circulate abundantly in the art world.

The third decision concerns the artist contract. Works created with the assistance of artificial intelligence raise unprecedented questions of intellectual property, reproducibility and conservation. The dealer representing an artist working with AI must ensure that the conditions of creation, reproduction and conservation of the work are clearly defined in the contract. Lisson Gallery in London was among the first to adapt its artist contracts to account for the specificities of digital and algorithmic works.

05Anticipating without rushing

The astute dealer in 2026 is the one who closely observes the evolution of artificial intelligence in the artistic field without rushing into a position they might later regret. Art market history teaches that new technologies always eventually find their place, but rarely in the way early enthusiasts predicted. Photography took decades before being fully recognised as an artistic medium in its own right. Video followed a similar path. Digital art, despite decades of existence, achieved significant market recognition only recently.

Artificial intelligence will probably follow a comparable trajectory: a phase of effervescence where everything seems possible, followed by a phase of settling where the most superficial practices disappear and the most rigorous artists emerge with works that durably inscribe themselves in art history. The dealer who can identify these rigorous artists amid the ambient noise possesses a considerable competitive advantage.

06The question of rights and transparency

One of the most pressing issues for the dealer interested in art involving artificial intelligence concerns copyright and transparency towards collectors. Generative AI models are trained on millions of images, many of which are works protected by copyright. Legal proceedings are under way in several countries to determine whether this use constitutes a violation of the rights of artists whose works served as training data. The dealer must be informed about these legal debates and able to explain their implications to collectors.

Transparency about the creative process is another imperative. The collector purchasing a work involving artificial intelligence has the right to know to what extent AI contributed to the final result. Some artists use AI as a starting point for extensive manual transformation, while others give it a more central role. The dealer must be able to document this chain of creation and communicate it with precision. Hauser and Wirth, which represents artists working with varied technologies, has developed detailed technical sheets that accompany works and explain the tools used, the artist's interventions and the conservation conditions specific to digital supports.

The secondary market for AI art also raises questions the dealer must anticipate. How does one resell a work whose creation process partly relies on an AI model that may have evolved or ceased to exist in the interim? How does one guarantee the authenticity of a digital work that is infinitely reproducible? These questions, still largely open, oblige the dealer to develop legal and technical expertise that exceeds the traditional competences of the profession. Galleries that invest in this expertise now will be best positioned when the AI art market reaches maturity.

Artedusa offers its partner galleries an international showcase to present these discoveries to a public of collectors attentive to the evolution of contemporary creation, enabling dealers to document their artists' creative processes and communicate transparently about the technologies involved in the production of works.