The headlines of recent years haven't always been kind to artificial intelligence, which is often painted as a looming threat to creativity – hovering on the margins, ready to swoop in and steal jobs. A sampling of stories seems to spell certain doom:
“Even the Mad Men could be replaced by machines someday”
“AI and creatives' complicity in their own extinction”
“If robots make art, what's left for us?”
However, despite its negative press, conversations about machine learning and AI are very much at the forefront of the creative industry. That means everyone from major ad networks – such as Publicis, which cut its entire marketing budget for a year to develop its AI platform Marcel – to marketers and designers concerned that their craft is under threat. Amongst the worry, there are those beginning to explore its vast potential as a creative collaborator.
“We thought [our employees] would be nervous about it: Is this going to kill off creative?” says Sam Ellis, Business Director of Innovation at M&C Saatchi, which has experimented with optimizing poster campaigns using machine learning. “What they started to realize is that it could be really, really useful based on its insight.”
It's an opinion shared by Sampo Kaasila, Monotype's Research and Development Director for AI, Machine Learning and Deep Learning, who sees AI as more of a partner than a competitor. Much as digital tools have removed some of the legwork graphic designers used to have to do (think of the painstaking process of using Letraset to make lettering), AI has the potential to do the same.
“I see a world of design where humans work side by side with AI,” says Kaasila. “I see them becoming an assistant, with the combination of human and AI more powerful than either of them alone. And it will become more and more natural. People don't think twice nowadays about using Alexa or Siri, and that's deep learning behind the scenes.”
I see [AI] becoming an assistant, with the combination of human and AI more powerful than either of them alone.
Bo Hellberg, an adland veteran, and founder of Kaiser, a creative consultancy for machine learning and innovation, also sees machine learning as a future partner.
“It will enhance how we conceive, ideate, make and produce for any media or format,” he says. “It won’t take anything away. It will most definitely give us more options, variations, iterations and opportunities to explore more expressions.”
And while significant changes to some industries seem more imminent, there's an understanding that AI isn't quite ready to cannibalize creativity yet. The limitations of machine learning mean that computers can only make work based on the data they've been fed, rather than drawing on the full range of human experience – as a result, the output is still limited. Certain complex tasks currently handled by designers and creatives (think designing typefaces, responding to a detailed client brief, or projecting a certain image or feeling) are still outside the realm of possibility for AI, for now at least.
“AI doesn't do things on its own,” says Joe Stanhope, vice president and principal analyst at Forrester. “There isn't a HAL 9000 you can go to and say, 'Hey can you put together this campaign for me?' It doesn't exist, and it's going to be a long time, if ever, until it does.”
So until AI is ready to produce its first Leonardo da Vinci, it's clear that its potential mostly lies in easing the creative process. It can carry out deeper research than a human ever could, helping to understand demographics and their preferences more thoroughly, and make recommendations accordingly, or work out what's popular and nudge designers to explore less 'obvious' directions. Creative work could become an increasingly hybrid process.
“We are basically building tools that are non-traditional and AI-based, that allow designers to do things they couldn't do before,” says Kaasila. “There are ways it can augment products, and the usability of products. No human can know about 133,000 fonts, but the machine can, so it can suggest. But you make the ultimate design decisions.”
No human can know about 133,000 fonts, but the machine can, so it can suggest. But you make the ultimate design decisions.
The presence of AI could also, ironically, push the creative world to think more independently. Creative Review editor, Patrick Burgoyne sees potential for AI to force designers out of a permanent feedback loop of recycled inspiration.
“What these systems are not good at is difference, idiosyncrasy or seemingly illogical departures from the norm,” he writes. “Creativity in other words.”
With all this potential, there remains plenty for the creative industry to learn. Institute of Practitioners in Advertising (IPA) president, and CHI & Partners chief executive, Sarah Golding has already announced plans to roll out an IPA training programme to teach creatives, with an agenda that's heavily focused on working with, rather than against, AI.
“The machines will be our new colleagues,” she said in her inaugural speech at the IPA Members' Lunch last year. “And they will be the smartest, fastest, most challenging colleagues any of us has ever had. But in order to make even more creative magic we must embrace the machines, enjoy the machines and trust the machines.”
What Golding's speech makes clear is that the creative industry can't just rely on investing money in the problem, but must also overcome its resistance – understanding how to work with this technology as a new part of their toolbox. Kaiser founder Hellberg also believes in order for us to make the most of AI creatives need to be prepared to spend time with it.
“Although there is this sinister vision of a future full of unaided automation, the reality will be that we need to pump in a lot of energy, insight and creativity into working with machine learning,” he says. “Communication will still need creativity to shape voice and values.”
“It may be possible for the machine to present a new creative idea you would never have thought of otherwise, or come up with things that don't exist in the real world,” adds Kaasila. “We think machines can't do creative work, but I'm not sure that's a hard boundary there. I think they can enhance our creativity.”