THINK art, think Da Vinci, or Michelangelo, Warhol or Hirst. Artificial intelligence? Not so much. But the latest innovation in art happens to be the world’s first portrait-drawing droid, which takes on the guise of a human female artist and blurs the boundary between machine and creativity. With her own exhibition at the University of Oxford, the ultra-realistic humanoid, called Ai-Da, is fast becoming the leading face of the AI art movement. But is she any good?
The first thing you notice when meeting Ai-Da is how eerily lifelike she is, with her humanoid head, neck and full head of hair. In fact, the blank expression plastered across her occasionally blinking face is unsettling, particularly when coupled with her stalker-like ability to follow you across a room.
I soon discover her movement is actually on rotation and controlled by a human, so the stare down is just a coincidence. Phew. In her white blouse with paint splotches, she also looks every inch the artist. But her robot roots are obvious, thanks to her openly exposed mechanical limbs that are capable of grasping a pen or pencil.
Ai-Da has been named after the world’s first female computer programmer, Ada Lovelace. She is the brainchild of gallery director Aidan Meller, who calls her ‘an incredibly sophisticated robot’ (she also has facial recognition technology) and claims she’s the first ultra-realistic robot capable of physically drawing people.
So how does a robot draw? Ai-Da has an in-built camera in her eye to identify, analyse and capture images in front of her, which are then interpreted and processed by machine learning. The data that process generates is subsequently fed into a series of algorithms that send coordinates to her robotic arm to start sketching. Our demo takes 45 minutes.
‘It varies each time because the algorithm interrogates the computer vision in different ways,’ Meller says.
Initially, it’s hard not to scoff at Ai-da’s attempt at a portrait of Meller, which looks more akin to the kind of thing a five-year-old would draw — or a piece of evidence used in a court of law to prosecute a serial killer. Even Meller calls her drawings ‘slightly unnerving’ but goes on to say that the aim of the artwork is not to produce pretty pictures for your living room.
‘She needed to have her own unique drawing style that would produce something original and unpredictable and offer another way of looking at the world,’ says Meller. ‘We could have got Ai-Da to do really beautifully finished drawings but she is not an expensive printer. She is fully algorithmic and fully creative. You don’t know what she is going to do.’
Meller may be the brainchild behind Ai-Da but it took a collaborative effort to bring her to life. She was designed and built by a team of engineers from robotics company Engineered Arts, while her artistic ability and algorithms have been developed by students at the University of Oxford and Leeds. The decision to make her a female robot also stems from the underrepresentation of women in the art world, a voice that according to Meller ‘is needed more now than ever’.
The art on show at her sold-out exhibition features Ai-Da’s pencil drawings, including sketches of AI theorist Alan Turing and Ada Lovelace, alongside paintings done by human artists that have been rendered from the coordinates taken from her drawings and put into various different algorithms to produce, for example, an oil painting entitled Bee-Shattered Space. Meller describes it as ‘purposeful experimentation’ into the ways humans and technology can collaborate to create art and hopes it will generate discussion. (see box on the right).
And if you’re wondering how successful she’s been so far, well, Ai-Da has already sold over a million pounds’ worth of artwork. Wowsers.
Rise of the AI artists
AI is coming to the art world in other ways too. Christie’s made history back in October with a distorted portrait generated by artificial intelligence of a gentleman dressed in black. It was the first algorithm-generated print to go under the hammer. Despite boasting the kind of visuals that wouldn’t look out of place in an Andy Muschietti horror film, it sold for $432,500 (£386,884) — over 40 times Christie’s initial estimate and more than double the price of the Andy Warhol print and Roy Lichtenstein bronze sculpture on display beside it.
In March, Sotheby’s auctioned a piece of artwork that came straight from the mind of a machine for £40,000. Titled Memories Of Passersby I, it consists of two screens generating never-ending male and female portraits. Most recently, the Barbican opened its doors to an exhibition designed to explore our relationship with AI, complete with an AI bartender added for good measure.
The AI art market is estimated to be worth more than £77.7million and has the potential to expand to ten times that size in the next three to five years, according to Research And Markets. But will AI artwork revolutionise the art world or destroy it?
Ai-Da’s artwork will be shown at the Unsecured Features exhibition at The Barn Gallery, St John’s College, University of Oxford until July 6
BOT matrix more impressive humanoids
The Kodomoroid TV presenter
Fresh from Japan, the Kodomoroid bot speaks a number of languages and can perform simple tasks, such as reading the news and giving weather updates, while motors embedded in her face give her realistic facial expressions. She works at Tokyo’s Museum of Emerging Science & Innovation and provides directions.
Loaded with over 160 motors, ‘the most advanced humanoid robot yet’ replicates the human body as closely as tech allows, complete with actuators that mimic the contraction of muscles. That enables it to do push-ups, backbends, pull-ups and even sweat, which could help inform the design of robotic limbs.
If there was an award for creepiest bot, this animatronic baby would be a frontrunner. Created by a special effects company for an exhibition at the London Science Museum, the realistic baby was designed to breathe as well as move its legs and arms, and could be used for medical training and parenting practice.
Sophia, designed to look like Audrey Hepburn, is the world’s first humanoid robot with a nationality after being granted Saudi Arabian citizenship in 2017. She can make up to 50 facial expressions, expresses feelings, her AI revolves around human values and she has appeared publicly to speak about women’s rights issues.
To join the dots between robotics and AI, Meller first thought about what all the great artists have in common.
‘They captured the zeitgeist,’ he says. ‘They touched on a nerve that reflected the undercurrent of their time.’
Meller, an art dealer for 20 years, concluded that in the next few years the zeitgeist underpinning almost every new innovation is AI. So he created a robot that could bring art into the world while facilitating a dialogue about AI’s potential.
‘The artwork [above] is probing into our world from a non-human perspective,’ he says, ‘and that creates a brand new voice that is not representative of the art world.’