What AI can teach us about generating ideas
The tech sector is spending an awful lot of energy and resources developing Artificial Intelligence and the media are lapping up this brave new world. We’re seeing self-driving cars, neural networks and the emergence of robots. Which, considering humanity’s track record, will begin their lives as either military or sexual machines (or in the worst/best of cases, both).
All this energy has been spent breaking down how humans function into a world of algorithms and functions (which thankfully you’ll need to know nothing about to read this article). But what if we reverse engineered insights on how intelligent beings built of ones and zeros make decisions and brought it into our creative practice? Read on:
Inputs = Outputs
There’s a wonderful parable about a military computer that was taught to identify tanks by being fed aerial photographs and being told if there was a tank in the photo. After a great many days, photos (and coffees), it was finally ready to be deployed in the field with an impressive 99% accuracy rate. The team eagerly awaited the photos from a real warzone and fed them to the intelligent machine overnight.
In the morning the team came face to face with an enemy force of epic proportion (a tank in every photo), which the computer had totally failed to identify. It turned out the photos they had trained the machine on had been taken on sunny days, when in fact, the ones from the warzone were taken on a cloudy day.
There’s a clear moral to this story. What you experience directly influences what you create. If you experience the same thing every day, your output will be stale. You need to make sure you balance out experiences and continually push for new ones (e.g. don’t read what others read or else you’ll just know what they know).
It never works the first time around
It’s easy to be wowed by the successes of humanity. Don’t be fooled by media hype, things don’t ever work the first time round. Asimo, Honda’s bipedal robot which made the headlines playing football with Obama, is the by-product of 20 years of development and countless iterations. Being a developer teaches you this the hard way by forcing a version number on your work. (The company I founded is now on version 175, and will probably be on 176 by the time you read this.)
Nothing comes easy. But nothing that comes easy is of value. Don’t give up — build, measure, and iterate. Your first work won’t be good. Neither will your second. But as long as your second is better than your first. You’re going in the right direction.
Complexity is layer, upon layer of Simplicity
We’re not far from being able to simulate the human mind. And it’s coming sooner than many people would be happy with. Every major Silicon Valley Company is pushing into understanding natural language (Alexa, Ok Google, Siri…). We’ve got robots walking, crawling and swimming — and we’re also pushing them into creating. But if you were to crack open a machine and view the code being executed, you would boil down the whole machine to ones and zeros. These achievements are extremely complex as a whole system but — at their very core — as simple as your bedroom’s on/off switches.
If you’re going to do complex projects, don’t ever take them head on. You’ll get disappointed at the lack of progress and give up. Make sure to break down large workloads into smaller more manageable tasks, which you’ll probably need to break down even more. If you can’t finish the task in a couple of days (hours would be even better), you won’t get it done. This also allows you to assess some of the moving parts as you go before putting them together, giving you a greater chance of success.
Failure is a part of Success
In the past couple of months, public interest in self-driving cars has spiked. And so it should; this is a major evolution not just for transport and logistics, but for mankind. Imagine a world where transportation is no longer an issue. You, your cat and your packages can go from A to B without any stress. This will also give us a chance to reboot the anomalies of the old world. No more driving on different sides of the road, no need for traffic lights, getting stuck in traffic jams will only ever be referred to in film (the brilliant Falling Down with Michael Douglas comes to mind).
But this panacea won’t come easy. There’s nothing like a grim accident hitting the headlines to put the brakes on progress (no pun intended). Of course, it’s horrifying and the loss of life is very sad. I would never excuse it. But lives are lost on the road all the time (1,775 in the UK alone in 2014; with a further 22,800 serious injuries), regardless of the driver being a person or an AI entity. In the long term, this technology will save 1,000’s of lives but we’ll have to cope with the occasional catastrophic failure.
But I’m digressing; the importance is that failure is a by-product of trying to do something innovative and different. You’ll fail and that’s ok. Accept it, learn from it, dust yourself off and continue on your journey.
Perspective is Everything
How you solve issues is very much based on how you define the problems. Coming back to automated driving, you could define the issue by breaking down the many rules of the highway code, add the complexity of a human’s capability of making decisions whilst driving and the systems available to the car. Or you could take the advice I got from an Egyptian friend of mine whilst driving in Cairo, “Don’t hit anything”.
It’s easy to be bogged down with the details; but if you were to boil down what you were to achieve, what would it look like? Understanding that doesn’t just allow you to establish if you are successful in what you are doing, but gives you space to innovate when you come across a problem.
People want to believe
Whilst in Japan in 2011, I passed by a store selling Sony’s now defunct Aibo dog — a quaint-looking piece of soft plastic that looks vaguely like your favourite canine (without the late walks and mess that goes with them). Seeing it at first didn’t convince me. It’s a hairless, eyeless object. Then it moved and started interacting with those around it. Children and old ladies promptly started petting it and speaking to it.
Years later, whilst giving my TEDx talk, I showcased another canine robot, this time built by Boston Dynamics. A much more expensive offering built of military grade materials, with 4 legs, a body and no head. The video showcases the incredible agility of the machine, tackling hills, steps and running around its environment. Then a man jumps out and drop kicks the machine from the side, prompting it to stabilise itself. The room gasped in horror, a reaction which would have been equivalent to that gentleman tripping over a Chihuahua at an uptown café, not someone hitting their computer out of frustration.
These examples prove humans favour suspending reality in favour of their imaginations. As a creative, you’re able to tap into people’s imaginations — so it’s great news their imaginations are on by default. The import thing is to make sure you don’t betray that trust, since it will be impossible to recover it afterwards (ever seen a film where you became aware of the people sitting around you instead of being wooed by the story. That’s when they lost your trust).
API’s / Open Source
If I were to build a robot today, I would have two options.
I could start from scratch; invent the eye, perfect a robotic arm, establish the best way to balance the unit whilst walking. Think of everything you do daily, without thinking — I would have to recreate all of that. Then I would need to spend years testing it all. Alternatively, I could reuse people’s existing work and build on top of that. Different companies are becoming experts in different fields. Some are working on the brain, others are working on the body. The only way technology is able to develop so quickly is through collaboration, communication and open systems.
Creativity is inherently collaborative. Don’t be ashamed to ask how others achieved what they did. And be kind with your time (don’t think you can protect your practice through secrecy, the internet killed that in the 90’s). You may learn something new and build the networks that will allow you to succeed in your craft. Imagine what you could achieve if knowledge wasn’t an issue?
Conclusion
Artificial Intelligence is well on the way to conquering the jobs available today. It started with factories a couple of decades ago and will extend to all corners of society (I would recommend professional drivers start diversifying their skillsets).
By the end of this century, machines will perform better than us at almost everything we do. Machines will even make machines, closing us completely out of the loop.
Of course, this comes across as scary (the movie industry has responded with Terminator / Ex Machina / Humans and now Westworld). Luckily for us, creativity will be our last bastion. So we best start getting better at it so we can keep our competitive advantage.
Guy Armitage
Founder of Zealous, an online creative network allowing creative talent to be sourced through open calls. He previously founded Bright-Creations in Cairo, Egypt; and was in charge of the FTSE quarterly review change for the London Stock Exchange. Guy was featured on the world-changing potential of creativity at TEDx and in Forbes; as well as winner of the Tate Digital Art Hackathon.
