Computational Creativity: Can AI Programmes Become Artists?
Ai.Da Self-portrait I by Ai.Da, 2019, pen on paper, 29 x 41.5 cm, © the artist | Image Via Time Magazine
‘Art’ has been defined as the expression or application of human creative skill and imagination to a medium, through which the artist can make visible their internal view of the world to an audience.
It is necessarily and fundamentally an interaction between human agents: a transference of views and experiences from one mind to another. The advent of Artificial Intelligence (AI) technologies has begun to destabilise this notion, questioning and subverting the boundaries of art-making and artists. For those who think of ‘art making’ as a distinctly and wholly human capacity, art-producing AI technologies are a cause for concern. But should we fear the works of these computer artists?
Artificial Intelligence has developed rapidly over the last decades with increasingly more interest from technological companies in the replication and emulation of human creativity using trained algorithms. These algorithms collate and assess data about human behaviour and reproduce them with uncannily human-like abilities. One such artistic project is Ai.Da, hailed as the world’s first ultra-realistic robot artist. The computer brain has been personified in a robotic body and a woman’s head, further blurring the line between human and computer forms. Ai.Da is a performance artist whose media are varied, ranging from drawings to collaborative painting and sculpture. Her creators acknowledge that she is not alive, but they view her as possessing an artistic persona that we respond to. Ai.Da produces artworks that are exhibited in museums and galleries worldwide under ‘her’ artistic license, most recently in Oxford in 2019, and like other artists, her exhibitions were also halted by the global pandemic this year. At the same time, it is difficult to pinpoint what exactly is interesting about Ai.Da as a phenomenon. Do viewers respond to the art that she produces, or do we primarily react to the questions of human-machine interactions that her oeuvre poses?
Consciousness. In short, the interaction of sitter and painter has been entirely outsourced to the creative liberty of an algorithm. Such human-machine relationships become even murkier in projects such as The Next Rembrandt, developed by Microsoft and ING alongside the Delft University of Technology and the Mauritshuis Museum. The premise of the project was an attempt to ‘bring back’ Rembrandt van Rijn, and to create a completely new piece of art in exactly the same manner and style as the great master. As such, the aim was to resurrect Rembrandt’s artistry and hand with computer software. The project began with a thorough analysis of all of Rembrandt’s paintings to collect data about his subjects, his handling of the brush and the pigments he used.
Shattered Space: Inua Apis-Ipsa by Ai.Da, 2019, Original oil and giclée on canvas, 99 x 99cm, © the artist | Image via AKG
The Next Rembrandt by The Next Rembrandt Project, 2016, © The Next Rembrandt
From this mass of information, the algorithm determined that the ‘next Rembrandt’ would be a portrait of a Caucasian male with facial hair, between the age of 30-40 years and wearing dark clothing, a collar and a hat, facing to the right. The next step was to extract data-information about Rembrandt’s paintings that matched the specification. Although the finished product is a collection of details that are typical to Rembrandt, the artwork is an 3D-printed original piece. It cannot be attributed to Rembrandt despite the fact that it is an exact mathematical approximation of what an original Rembrandt portrait would be. Even more than the problem of attribution however, it also challenges the foundations of portraiture as a genre. The result is a painting of a man who does not exist, created by an entity which does not possess.
Faceless Portrait of a Merchant by AICAN and Ahmed Elgammal, 2019, Creative Adversarial Network print on canvas, 76.2 x 76.2 cm, © the artists | Image Via HG Contemporary
Although the existence of AI artists challenges the notion of the uniqueness of human creativity, it can also become a new mode of artistic expression in which the computer enables human development. In 2019, the HG Contemporary Gallery in New York organised an exhibition titled Faceless Portraits Transcending Time which was a collaboration between Dr Ahmed Elgammal and an AI programme named AICAN. Similarly to Ai.Da’s exhibitions, the pieces were all digitally generated using AICAN, but the emphasis was placed on the AI as a programmable medium in which the machine can only exist in collaboration with human creativity, not as a separate entity from it. As such, even though Dr Elgammal is a computer scientist and the physical works were created by AICAN, the artistic value was created through their interaction. In this instance, the machine is not anthropomorphised but treated as a mathematically intelligent tool for human expression. It replicates aspects of human creativity but does not have the capacity for genuine independent creation.
Portrait of a Scientist by AICAN and Ahmed Elgammal, 2018, Creative Adversarial Network print on canvas, 76.2 x 76.2 cm, © the artists | Image Via HG Contemporary
It is difficult to judge how AI artists and art might develop in the near future. AI technologies could be seen as today’s avant-garde which disrupts the pre-existing norm and portends an imminent paradigm shift in human-machine relations. Will we see AI artists be subsumed into the canon, or will they fade in time as a passing fad of a technologically-crazed generation?
Thank you for reading,
Aniko Petri & Cluster Team.