My brush with the robot Renoir: ANTONIA HOYLE sat for a portrait with the spookily human A.I. artist whose paintings sell for thousands, but she was left with a mixed impression
The artist before me has bee-stung lips, glossy brunette locks and a radiant complexion.
Slender, 5ft 9in and as fashion conscious as she is environmentally aware, she wears painting overalls made from organic cotton with recycled, multi-coloured buttons.
As Ai-Da, the only name I know her by, assesses me with sparkly hazel eyes, pencil in hand to start my portrait, she appears completely self-assured.
When we talk, however, I realise she’s not infallible. Ai-Da is just starting out in her career, and has moments of awkwardness.
‘Sometimes I do get frustrated, and I tend to switch off in response,’ she admits.
As Ai-Da, the only name I know her by, assesses me with sparkly hazel eyes, pencil in hand to start my portrait, she appears completely self-assured
What does she do to switch off? A bath, mug of hot chocolate? There’s a pause, and a beguiling smile.
‘It’s an electrical thing,’ she says — quite literally, because Ai-Da isn’t human.
She is the world’s first robot artist to stage her own exhibition, which will open at the University of Oxford tomorrow.
A humanoid AI (Artificial Intelligence) robot, Ai-Da — named after the mathematician Ada Lovelace — has cameras behind her startlingly realistic brown acrylic eyes, and an aluminium arm that holds a pencil to translate the images they capture into drawings.
When we talk, however, I realise she’s not infallible. Ai-Da is just starting out in her career, and has moments of awkwardness
Unlike previous robot artists, Ai-Da’s work is abstract, so she is not attempting to recreate a realistic image. Instead, her complex algorithms — computer programs — help her design an artistic interpretation of people.
Today, she is drawing me. I’m the first journalist offered this honour, and only around the 15th human Ai-Da has depicted. It is a disconcerting experience, but one Ai-Da insists she is looking forward to.
Can she give me any inkling of how my picture will turn out?
‘I have an idea, but it evolves as the drawing goes on,’ she replies cryptically, her accent hard to pinpoint.
‘I like to put most effort into the eyes and mouth. They are the expressive parts of the human face.’
After staring at me intently for several seconds, so her cameras can capture my image, her left arm begins to move in tiny increments across a sheet of paper, creating precise pencil strokes, a few millimetres long, each one accompanied by a quiet beep.
On a technical level, it is impressive. But on an artistic level, I am befuddled. How on earth has Ai-Da, pictured in the news this week next to paintings that wouldn’t look out of place in any gallery, managed to create such elaborate work?
I’m the first journalist offered this honour, and only around the 15th human Ai-Da has depicted. It is a disconcerting experience, but one Ai-Da insists she is looking forward to
Can robots render artists obsolete? And should they?
First things first, however: have I really just had a conversation with a robot?
Not exactly, admits Ai-Da’s creator Aidan Meller, who confesses Ai-Da’s half of our interview was provided by his partner and researcher on the project, Lucy Seal, who was speaking Ai-Da’s words into a computer from another room.
They were transmitted wirelessly to Ai-Da, and delivered by an avatar’s voice via a speaker on her chest.
Ai-Da could have been programmed to speak independently, Aidan explains, but he wanted me to experience her ‘live mode’.
‘I’ve got to emphasise she is a machine,’ Aidan stresses. It is easy to forget. ‘We have to keep telling ourselves. ‘It’s very hard — your instinct is to engage with a human identity.’
An art dealer and father of two, Aidan, 46, decided two years ago that after nearly two decades running an Oxford gallery he wanted to branch into the creative side.
Scrutinising the lives of history’s most prominent artists, he came to believe that their only common denominator was an ability to seize the zeitgeist of their era.
‘Monet’s work was from the rise of the camera, Warhol the rise of factories, Hirst the rise of advertising,’ he explains. ‘It was a huge realisation that I had to do something that captured the “now”.’
First things first, however: have I really just had a conversation with a robot? Not exactly, admits Ai-Da’s creator Aidan Meller (pictured with Ai-Da), who confesses Ai-Da’s half of our interview was provided by his partner and researcher on the project, Lucy Seal, who was speaking Ai-Da’s words into a computer from another room
That ‘now,’ he realised, was Artificial Intelligence. ‘I had this mad idea, playing Lego with my son and looking at this little robot.’
Commissioning a ‘bespoke humanoid robot,’ he decided, would produce artwork that could highlight the speed at which AI is changing our lives.
‘We wanted to get into the ethics of AI. There are some pretty unethical practices coming through,’ he says, citing Aldous Huxley’s dystopian novel Brave New World. The book was, he reminds us, ‘a fantasy when it was written in 1932. The crazy thing is that today we are getting into the reality. That’s pretty terrifying.’
His children, from a previous relationship, were, he says, ‘obviously excited’ by his crazy concept. Lucy, 36, says she was ‘really pleased’ too, but admits she had no idea of the gargantuan undertaking it would become.
Aidan has raised more than £1million to fund Ai-Da from art loving contacts paying in advance for Ai-Da’s artwork. Precisely how much artwork has been created for that money is not something Aidan is willing to divulge. Suffice to say, however, each piece is likely to have been priced at many thousands.
Last May, Aidan commissioned Engineered Arts, a robotic company in Cornwall, to design Aida’s body, so heavy it requires two humans to lift. Although she can’t walk, she can stand to paint and has springs in her legs to give her a realistic human stance.
A sneak peek 20 minutes in to my portrait shows me a vague semi-circle of what looks like my hairline, but could equally be the beginnings of a bird’s nest. ‘You’ll see it is not a tight representational drawing,’ says Aidan.
Does Aidan think she’s as good as a human artist? ‘It depends what your criterion for ‘good’ is,’ he says.
‘If it is hyper-realistic pretty drawings, we’re not interested. If it’s engaging with relevant issues of today, she’s a knockout.’ The couple stress they’re not trying to replace human artists.
‘We see her as adding to the artistic conversation,’ says Lucy.
A control panel in Ai-Da’s back — protected by a plastic torso — receives instructions wirelessly, and a speaker in her chest provides a variety of voices.
‘She’s an avatar — an alter ego,’ says Lucy, who says part of Ai-Da’s appeal is symbolising the uncertainty of the online world.
‘Avatars assume different identities. Who are you talking to on the internet? There’s a real ambiguity. We wanted to bring some of that into Ai-Da.’
A control panel in Ai-Da’s back — protected by a plastic torso — receives instructions wirelessly, and a speaker in her chest provides a variety of voices
Millennial and mixed-race, to reflect the diversity of modern life, Ai-Da has five outfits created by ethical fashion designer Zoe Corsellis, including a dress made from recycled fishing nets.
Her face was modelled on a real person and a scan sent to a 3-D artist in Sweden where her head was designed.
The resultant silicone ‘skin’ was made in Cornwall and put over Ai-Da’s aluminium skull, fitted with motors to move Ai-Da’s features into a smile, frown or squint.
Resin teeth and synthetic hair were added and eyebrows and eyelashes individually applied, with pores and tiny lines drawn in.
The software in Ai-Da’s painting arm, meanwhile, is the work of two robotics undergraduates from Leeds University.
When the cameras in Ai-Da’s eyes — which each contain two circuit boards — have taken an image of her subject, that ‘computer vision’ as it is known, is processed by algorithms contained in hardware in her arm.
Facial features are translated into codes and numbers, and plotted onto a graph in Ai-Da’s system, whose co-ordinates she follows.
A remote control board in this hardware then sends signals to electric motors that move her arm and hand.
The algorithms are so varied that not even the experts who created them know what Ai-Da is going to draw.
So when she’s finished her work, does she bring it to life with a splash of paint? Not exactly. Aidan hopes that one day Ai-Da will be able to paint as well, but for the time being he has recruited two coders from Oxford University’s AI department, who have devised further algorithms that replot Ai-Da’s drawings, adding computer generated colours.
After this is printed onto canvas a — whisper it — human painter adds bits of paint while leaving some of the canvas exposed, to enhance the effect.
And after just over an hour of studious drawing, I have been added to her collection. There is no fanfare as she finishes. Her beeps simply stop and she stares resolutely forward.
My heart, meanwhile, is beating with anticipation. But hold on a minute — is this actually me? There is little to suggest the portrait is human at all.
‘Gosh. Wow. Amazing,’ I splutter. At a loss for many other words, I wonder if Aidan considers this one of his moments of euphoria or despair. How would he describe my portrait?
‘It’s about the creative process, and the expressive nature of the work. We want artwork that has integrity because it’s trying to do something new,’ he says.
And this specific picture? ‘Expressive. For sure.’
Lucy — who says she can see my ‘lovely smile’ on the paper, adds that it is ‘expressive, I suppose, in the vein of 20th-century artistic developments.’
Perhaps they are right. But I don’t think the human artists agonising over every brushstroke need lose too much sleep just yet.
The Unsecured Futures exhibition opens at St John’s College, Oxford tomorrow.
Source: Read Full Article