Rise of the Creative Robots
Philip K. Dick may have wondered whether Androids dream of electric sheep, but perhaps a more fitting question is whether Androids ever dream of being MORE than electric sheep. Whether they will ever be able to tap into their very own electric creativity and walk their own paths.
Will we ever see an A.I. turn its hand to what we humans consider the finer, creative arts, the world of music, theatre, painting, comedy or even poetry?
Will we ever live in times where we can listen to symphonies composed by a Bach-Bot, an Amadeus Android, or marvel at the impressionist mystique of Monet-Machine; have our hearts stirred by the sensuous soliloquies of a cyborg Shakespeare?
Robot creativity is something humans have long both coveted and cowered from. We want cool, creative machines, but we're also terrified of them and the implications for ourselves. To date, however, it's been a bit of a moot point, as robot "creativity" seems rather hard to come by, anyway.
Sure, a cleverly designed machine can be programmed to simulate creativity, but does that count? Does good copying make good art? (FYI, Steve Jobs said it did!)
In 1997 the chess-playing computer Deep Blue beat world chess champion Garry Kimovich Kasparov. It was the first match Kasparov had ever lost and he put it down to feeling "spooked" by the machine, which he said had displayed "an unexpected creativity."
Of course, that creativity can be dismissed by mathematicians and A.I. experts as simply being able to calculate more permutations of moves, but isn't that the very essence of creativity? Seeing all kinds of different options that others haven't?
"Ok, so chess playing robots are pretty cool, but it's not the most creative thing I could ever hope to see a robot do," I hear you say. True. And to date, much developmental work in robotics is going into building machines capable of basic creative expression through natural language; bots that can understand what is being said to them, and in what context, so they can respond with something that makes sense. Because, without context, there can be no real creativity, or at least anyone to understand why a particular thing was creative. But that's complicated.
Even taking a step back from natural language to musical composition and analysis, robotic creativity is still in its infancy. Sure, you could say most music today sounds like it may as well have been composed by robots for all the monotony of it, but that might be selling robots short (Bieber-bot anyone?). After all, some robots have incredible musical depth.
For example, composer David Cope, who also happens to be a UC Santa Cruz professor, recently invented software that analyzes music and can find styles, patterns, rhythms and harmonies within each piece, which it can then flag as being signature to a particular composer. Something as individual as a fingerprint, apparently. The machine, "Experiments in Musical Intelligence" or EMI (Emmy if we want to personalize the machine) can, like the infinitely cute No. 5 from Shortcircuit, disassemble and reassemble pieces of music to make whole new pieces out them; pieces that even sound very much like they were written by the same composer. But is chopping something up and rearranging it inventive enough to pass our elitist standards of artistic creativity? Kanye West might not think so. Then again, Kanye doesn't seem to think much of human artists other than himself or Beyonce either! Imma let you finish, but...
Apparently, in most tests done on the subject of robot-made music, humans have not been able to tell the difference between the works of man and machine, and again, perhaps owing a little to its mathematical nature, music is well suited to automated composition.
What about art though? ART art... not musical art... but pictures, sculptures. There's actually an amusing short story by the literary godfather of robotics - science fiction writer Isaac Asimov - which describes a dinner party at a beautiful luxury home owned by a very wealthy hostess-with-the-mostest. The sumptuous house is filled with amazing sculptures of melting colors that leave the guests in awe. Subsequently, they find out the sculptures have been made by a faulty house robot, whose wiring is on the fritz. A guest - presumably a well-intentioned engineer - decides to fix the robot, and the hostess flips out. So much so, in fact that she kills the guest for meddling with her machine's maladjustment. A bit drastic perhaps - and altogether too much drama for a dinner party - but you get the picture. Asimov basically posits that only a robot on the blink would be capable of any real creativity.
Maybe.
There are, of course, robots today that can draw and create what most humans would consider to be quite palatable art, though these still operate within the confines of their programming. They don't paint out of "inspiration" or "soul" but because they have been given parameters and are able to execute on them. The results, however, are often unique and quite beautiful.
So is it just the intention, and the thought behind something that makes it creative, or the end result? After all, would we prefer something that was painted with heart but looks hideous, or something painted by machine which ends up looking fantastic and quite unlike anything else ever painted?
The key, say some roboticists, is that to be truly creative, an Android should be able to critique its own work. To learn and grow, to feed off of interactions and experience. That day seems rather a long way away still, but for now, at least, we have robots that can at least outsmart us at Jeopardy!
Whether we are based on carbon or on silicon makes no fundamental difference; we should each be treated with appropriate respect.