Tobias Revell on craft and AI: 'Machines might actually enhance our human-ness'

In his 1994 article The Technology of Enchantment and the Enchantment of Technology, social anthropologist Alfred Gell suggests that the power of great objects of human art, craft and ingenuity lies in their technical accomplishment. The profound spiritual resonance of great religious artworks, for example, is a result of them being incredible human achievements that the viewer might struggle to comprehend and so attribute to divine forces, just as the astonishing skill of a glassmaker, woodcarver, a potter or embroiderer might appear to be supernatural.

A similar process has been well documented when it comes to artificial intelligence and predictive systems: the apparently superhuman virtuosity of machines when playing ‘complex’ games such as chess or ‘Go’, or when recognising images and interacting with humans, imbues them with seemingly superhuman prowess. This is often exploited by tech companies, which draw on magical or mystical terminology in describing how they work. This is, perhaps, why engineers and tech companies are attracted to art: it allows them to activate enchantment as a way of demonstrating the power of AI. As Gell identified: if machines can make great art or craft they must surely meet, if not exceed, the human.

A recent example of the enchanting qualities of AI is Ai-Da, ‘the world’s first ultra-realistic artist robot’. Ai-Da is perhaps best known for its appearance at a House of Lords select committee hearing in October 2022, where it testified on the role of technology in the future of the arts in the UK. This year, Ai-Da designed a collection of ceramics for an exhibition at the London Design Biennale at Somerset House – each of the 3D-printed objects has the appearance of being individual, with the flourish of craft and decoration rather than the minimalist utilitarianism an audience might expect from objects designed by machine.

The robot’s website describes ‘her’ as ‘an artist-in-residence at Porthmeor Studios in St Ives, influenced by modernist design, including Bauhaus, the Omega Workshop and the Leach Pottery’. The objects themselves are intriguing, referencing the art of kintsugi with a dash of the extra terrestrial. Like much of the narrative around Ai-Da, the work is intentionally ambiguous, framed as ‘provocations’ that attempt to trouble the boundaries of human and machine creativity: why would a machine design ‘inefficient’ objects? How would a machine develop an aesthetic taste?

However, the success of Ai-Da, both as an iconographic object of discussion about AI in the press, as well as the future of the arts, is an intentional and well-crafted exercise in enchantment. Little information is ever given about the exact process or algorithm by which Ai-Da ‘creates’ art and the levels of human intervention and labour involved in programming this supposed spontaneous creativity. In fact, engineers working on the project have commented on how profoundly unsuitable Ai-Da is for artmaking. Its human form and visually pleasing creations obfuscate the technical reality of a team of engineers and backers who carefully choreograph the robot’s appearance and outputs to meet the expectations of a public trained on decades of science fiction about what robots and AI are ‘supposed’ to look, sound and feel like. It is no more an artist than a kiln, a microphone or a printer, but draws on the existing enchantments a general audience has about AI as somehow superhuman or gestalt.

Instead of being entranced by such ‘skills’, we should look at artists who seek to disenchant us of the dominant narratives of AI and enlighten us about alternative ways in which these machines might actually enhance our human-ness and extend our creative register.

Sougwen Chung’s improvisational performances and paintings such as Assembly Lines (2022) explore the intertwining of ‘biological and mechanical subjects’. Rather than exploiting science-fiction stereotypes, Chung interacts with the machines she calls her collaborators on their own terms. She moves about the floor at their level, reading the movements of the robot arms and responds to, rather than directs, the artistic product. Where Ai-Da is a carefully choreographed performance machine that attempts to convince an audience of its individual accomplishments as an artist through relatively consumable forms (a humanoid female robot creating paintings and ceramics), Chung’s work and the work of other critical practitioners attempts to bring the audience closer to the machine’s perspective by playing with movements, shapes and interactions particular to it.

We might also compare Ai-Da’s ceramics collection with Spawns, a collaboration between design studio Oio and jewellery designer Giosampietro. In this project, a machine-learning software was trained on thousands of images of spoons to generate 100 unique new designs that were then crafted by the studios – a process of back and forth they refer to as ‘artisanal intelligence’. In this process, the role of the human and the machine is made clear and transparent and creates unique opportunities for new forms of craft to emerge. The machine is able to process its perception of the object at a vast scale and suggest novel forms for the human maker.

The work Ai-Da produces tell us very little about humans, machines and craft because (even if we buy the story that Ai-Da is somehow autonomous and devoid of human input) it aspires to simply replicate existing human creativity in machine shape. However, practices like Oio’s, Chung’s and many others explore how we might better understand what it means for humans to make, when we approach AI as it is, resisting the urge to dress it up as something we are more comfortable with.