Artificial intelligence is coming for your jobs

Bo Hellberg is an Executive Creative Director and leads 'An Introduction to Creative Digital Advertising' Masterclass - a must for anyone who needs to swot up on their creative digital skills.

Taken from his 2015 Digital Shoreditch session, in this article, Bo offers some valuable creative artificial intelligence insights and looks at how these will affect the advertising and design industry in the future. 

Human beings are really slow. 

Not only are machines faster performing, but they also have an advantage when it comes from generational shift, ‘evolution’ if you like. With machines, information, intellect and experience is immediately passed on to the next generation. This doesn't happen with humans.

Take for example Salvador Dali, it took him a lifetime to achieve that level of skill. If a baby was born with that already, imagine what he could achieve in his life. And that’s what I find fascinating about artificial intelligence, it offers us that possibility.

Let’s talk about computer programmes capable of human level creativity. This means understanding human level creativity – and formulating an algorithmic perspective on creative behaviour. This is really hard for machines, because humans make no sense when we’re being creative. 

We use upwards of thirty cognitive strategies when we approach a creative problem. We mash things up. We create harmony, discord, symmetry, asymmetry, balance and imbalance deliberately throughout the creative process. 

It’s not logical, and therefore difficult for machines.

But machines are on the threshold of being able to read and understand semantic content. Right now machines only understand about 25% of what’s written. Very soon they’ll be able to read and understand everything that’s ever been written. All of a sudden we’ll be able to have a dialogue with machines.

The forerunner of this was IBM’s machine Watson. It does natural language really well, as it demonstrates here:

Machines can’t see very well. Yet. They lack focus; they don’t know what’s important in an image. And if they haven’t seen it before, they don’t know what it is.

In this test a number of computers were given an image and asked what they saw:

We can see here that AI machines are quite agnostic; they give options because they don’t know if they’re correct.

That’s exactly the question asked by the WHIM (What-If Machine) project – a collaboration between five universities in Europe. The explicit aim of this is “to build a software system able to invent, evaluate and present fictional ideas with real cultural value for artefacts such as stories, jokes, films, paintings and advertisements.”

If you work in the areas of social media and content, you’re screwed. Machines can come up with this stuff so much faster than you. They can produce content, come up with relevant ideas and personalise it quicker and more efficiently that we can. The kind of content we’re producing in these areas today is not that complicated. Look at how Nike are doing this:

Retail environments are already going more digital, and as their content needs updating continuously, AI will do that. There’s lots of data involved, Ai is essentially a big data platform.

As for direct marketing, that’s a no-brainer. That’s where AI is going to go straight in.

So machines can produce storylines, metaphor and craft. But the thing that’s saving us at the moment is that machines don’t understand what good is… yet. They can carry out user testing, but they can’t judge themselves.

Imagine having an ideation partner, it comes up with combinations, plots, characters and variations. It’s not a replacement for the creative, but it could be an incredible partner for creative problems.