The bots are coming for creative jobs, too

We’ve grown used to the idea of robots automating complex, manual tasks - like assembling a car. But the new wave of smart software is going to automate work that involves communication, intuition and even creativity.

A friend of mine works in financial journalism. Many of his firm's publications are now written entirely by algorithms that trawl through financial data and produce reports and briefings.

There was a moment, recently, when he realised many of the ’subscribers’ to those financial reports were… computers. The computers ‘read’ the reports and used them to inform financial trades.

It’s not science fiction. We live in a world where computers are writing reports for other computers to read.

Watch ING’s short online documentary The Next Rembrandt and you’ll see a team asking the question: can we teach a computer to paint a completely new work of art that’s indistinguishable from Rembrandt’s own work? The result is perfect - from the styling of the subject to the brushstrokes on the canvas.

When creative people see computers generating art it makes them anxious. It’s too easy to fall back on the ‘there’s nothing like the human touch’ argument. But if the audience believes it was created by a person, then what’s the difference?

So there are two questions that creatives need to address.

First, is my job at risk? Second, is human creativity itself is at risk? What is the point of people if computers can create convincing, emotionally moving works of art?

Each new wave of technology displaces old jobs and creates new ones. So people don’t spend all day ploughing fields or building cars, because one person with a tractor or an automated production line can do it in a fraction of the time.

But while roles disappear, work and employment continues, and increases. The smart question is not ‘is my job at risk?’, but ‘how will work change?’. Which brings me to my second point.

Science fiction has created a myth that computer intelligence is ‘pure logic’. But that’s not true. Algorithms are written by people - and the choices and biases that person makes are encoded into it. If the person writing or training an algorithm makes bad choices then the algorithm will make bad choices, too.

In other words, an algorithm is the result of a creative act. Like any work of art, it’s the result of a point of view about the world. You can hear that in this video - the team chose Rembrandt as the artist, they chose to create a portrait, they chose that the subject should be a man.

A symphony written by an algorithm is the creative act of the person who wrote that algorithm - the symphony emerges from the musical ‘taste’ and ‘knowledge’ which the algorithm’s author chose for it.

A different person, in different circumstances would have made different choices and another set of symphonies would have emerged from the algorithm. Authorship and creativity are still relevant to algorithms.

The opportunity we have is to create algorithms that generate art and entertainment that are tuned, specifically, to an individual. And to do so in such a way that the result is still surprising and engaging and not simply about amplifying an individual’s preferences until they’re locked in an echo chamber of unchallenging, uninteresting content.

I know people who are doing exactly that - creating algorithms that have the capability to surprise and delight, while being tuned to the needs of individuals. The team at Spotify Discover Weekly, for instance, have created an algorithm, and a delightful service, from their remarkable perceptions about how people share music.

The challenge for creatives is to understand their creativity, and to teach algorithms to speak, movingly, to millions of people individually. That future isn’t scary. It’s incredibly exciting.

Giles founded cxpartners with Richard Caddick in 2004. He's author of 'Simple and usable' and an invited speaker at design conferences around the world.