Ai-Da: art or illusion?

‘Ai-Da: Portrait of the Robot’ is a small but intriguing exhibition at the Design Museum in London, featuring Ai-Da, the self-proclaimed ‘world’s first ultra-realistic artificial intelligence robot artist’.

She is appropriately named after the English mathematician Ada Lovelace (1815-1852), who published the first algorithm for use with Charles Babbage’s Analaytical Engine, one of the earliest mechanical forms of the computer.

Ai-Da is a step towards an AI future that some anticipate with excitement whilst others dread. She was created in 2019 by a team of computer programmers, artists, mathematicians and engineers led by researcher Lucy Seal and gallerist Aidan Meller (who is also Ai-Da’s art dealer; sales of her art works have already raised over $1 million).

She has realistic silicone skin designed in conjunction with the company that made robots for the TV series Westworld, cameras built into her eyes so she can make ‘eye contact’, a body which allows her to twist her head and neck, a robotic artist’s arm designed by undergraduates at Leeds University, and a variety of hairstyles – she looks like a cross between the female robot in Fritz Lang’s 1927 film Metropolis and Ryan Gosling’s holographic girlfriend in the film Bladerunner 2049. Robot she is, but in what way is she intelligent? And is she an artist?

When it comes to the relationship between artificial intelligence and human beings, there are two problems into which we fall; we can reduce the human to the machine or elevate the machine to the human. The first of these believes that below the surface, human beings are merely biological algorithmic engines, and therefore nothing more than sophisticated versions of Ai-Da. This certainly seems to be the direction taken by some parts of the exhibition which show how Ai-Da can simulate human speech and even create poetic-sounding sentences using sophisticated speech recognition and AI language algorithms.

The opposite tendency, that of elevating the machine to the human, is ably demonstrated by a short performance art video called Privacy, in which a ‘naked’ Ai-Da is dressed by members of the public using pieces of cut-up clothing. What interested me is that those dressing Ai-Da treated her in a human way even though she is a non-thinking, non-feeling robot. The first two participants, both women, began by restoring Ai-Da’s dignity by covering her chest and pelvic area. Some of the men taking part appeared awkward about having to dress a ‘woman’, while others talked to Ai-Da, showing her the pieces of material they had chosen, asking her opinion, and then thanking her when they had finished. What this demonstrates, as any observer of pets and their owners will know, is that humans are relational beings who have an inbuilt tendency to anthropomorphise the non-human. This impulse is almost impossible to resist when that non-human has a face, makes eye-contact, and can carry out rudimentary conversation. But is Ai-Da’s apparent humanness a reality or an illusion?

As I walked around the exhibition a second time, something very human did emerge, but it wasn’t in Ai-Da or in her art. As the museum display noted, researcher Samuel Dickens had created the algorithmic code which allowed her to draw, but had chosen to do this in such a way that she made portraits in ‘a fragmented style influenced by 20th-century [human] artists’. We were told that ‘the hand of art technician Suzie Emery’ had contributed to Ai-Da’s finished portraits and that this ‘blend of machine and human creativity manifests the fractured nature of postmodern identity’. I began to wonder, however, whether the human input in Ai-Da’s art reflected the truth that her creativity was an illusion, and that the real talent on display belonged to her human inventors and collaborators. This was confirmed when I read one of Ai-Da’s AI-generated poetic utterances: ‘Letters are not finished by halves; they stand as indivisible parts, though letters be incomplete, yet complete, letters be one.’ At first I marvelled at Ai-Da’s creative genius, but on further analysis I started to wonder what such a sentence actually meant. Then I realised that the real genius on display was not Ai-Da herself, but the computer programmers who had done such a good job of getting her to simulate something so close to human language generation.

The writer and apologist Ellis Potter defines art as ‘deliberate human action’. Art, he says, is a relationship not a commodity, a communication from human to human. If this is correct, then the real art in this exhibition is not Ai-Da’s portraits, but Ai-Da herself. Neither is Ai-Da the real artist on display. The true artists are the myriad of brilliantly creative human beings who collaborated to make Ai-Da. The illusion of this immensely stimulating exhibition is to suggest that Ai-Da is an artist and therefore approximate to a human. The reality, as in the Wizard of Oz, is that behind the ‘curtain’ of Ai-Da are human beings very different from algorithmic robots. Humanity is safe… at least for the time being.

‘Ai-Da: Portrait of the Robot’ is at the Design Museum, High Street Kensington, London, until 29 August and is free entry, though you need to book ahead.

James Paul is director of the English branch of L’Abri Fellowship, a Christian study centre and community which provides people with the time and space to think through their questions about the truth and rel-evance of Christian faith. See www.englishlabri.org

CORRECTION

In this column in the August issue of en the rating for the film Sound of Metal was listed as being an 18. This was incorrect, it is in fact a 15.