Imagine a world where your toddler’s scribblings, now so lovingly pinned to your refrigerator door, warts and all, are perfect masterpieces.
One in which your nappy-wearing little’un skips the crayons and jumps straight to Creative Cloud levels of professionalism. One where all personality, all growth, all progression and, ultimately, all creativity, has been stifled.
That’s a future that Google seems to be working towards with its AutoDraw AI experiment. Revealed back in April and showcased at the company’s annual Google I/O blowout last week, it uses machine learning to to turn your scrappy sketches into stroke-perfect cartoons. Here it is in action:
It’s incredibly impressive, working in near real time to translate your wobbly lines into identifiable forms. And I find it incredibly unsettling, too.
A canvas of convenience
Google’s aim here is laudable – by employing AI analysis, it wants to offer a convenient way for anyone, no matter their artistic capabilities, to put together pro-level doodles in seconds.
Can’t find the perfect birthday cake drawing to stick on your party invite? Use your mouse or smartphone screen to draw a few candle-like shapes on a blob, and watch it transform instantly into a Bake-Off beauty.
But it’s also shortsighted, and missing one of the fundamental points about art. Our mistakes, and the things we learn from them, are what make us human, our experiences and our response to them the greatest expression of ‘self’ that we have.
By trading our imperfections for convenience, we will deny ourselves the opportunity to present the idiosyncrasies that make us individuals. And, in an increasingly humanist world, individuality is the most important thing that we have.
What would Google’s AutoDraw have made of Picasso’s heavily-stylised art works? They’re clearly not what you’d call ‘accurate’ representations of his subject matter. But though The Weeping Woman looks chaotic rather than human, its abstraction presents a deeper insight into the suffering she feels.
However, with feelings themselves abstractions (you can’t pick up happiness, or physically kick loneliness) is a disembodied AI – itself an abstraction of human thought – better placed than even ourselves to recreate these concepts in a definable medium?
(As an aside, Picasso's own words on the painting are illuminating – he was not averse to seeing man in the machine.
Speaking of his mistress Dora Marr, his muse for the painting: “For years I've painted her in tortured forms, not through sadism, and not with pleasure, either; just obeying a vision that forced itself on me. It was the deep reality, not the superficial one […] Dora, for me, was always a weeping woman….And it's important, because women are suffering machines.“)
Perhaps the more pressing question, one that avoids the philosophical meanderings of whether an AI could ever be capable of true inspiration, is whether Picasso would ever have had the drive to create his greatest works if he could have offloaded the gruntwork to an AI helper.
If the first time pre-school Picasso picked up a pencil (or its 21st century equivalent, the digital stylus) and saw his unique linework, however amateur, transformed into a homogenised form, would he have accepted the convenience and simply moved on to another distraction? Is struggle an inherent part of finding inspiration?
What would the co-operative output of a master painter and networked brain have even looked like? If Picasso were working today, living alongside the likes of AutoDraw, would his journey from realism to abstract pioneer have ever happened?
Our world is becoming ever more automated, and at some point we will need to draw the battle lines as to where we accept handing responsibility over to an AI judge.
Today we may be happy to let our doodles be scrutinised by a machine, but will we feel the same when a computer brain judges other aspects of our lives and the way we express ourselves?
“To err is human; to forgive, divine”, wrote the poet and playwright Alexander Pope. When our autonomous overlord finally figures out that it’s more capable than us in almost every aspect, assessing our every “err”, we’d better hope we’ve programmed a benevolent god.
- Gerald Lynch is TechRadar’s resident futurist. His weekly Future Gazing column casts a critical eye over the technologies and trends that are set to shape our world, bringing back to today a glimpse of tomorrow in the boot of his Delorean.