AI as Creative Director? Innovation Social & Alexa Stop talk Computational Creativity
Can computers genuinely be creative? Can they move beyond the role of Creative Assistant, and become Creative Directors? Episode 016 of Alexa Stop is a discussion about computational creativity, and how it could shape and impact the future of the advertising industry.
While it appears to be a rabbit hole of opinions, and crossings over of what computational creativity actually means and the direction it will take, the bottom line is that the future is quite literally in the balance… of data and interpretation. There is a future of creativity that isn’t completely owned by an algorithm.
Jim is joined in the studio by Lawrence Weber Managing Partner at Karmarama, Rachel Falconer Head of Goldsmiths Digital Studios, Alex Hobhouse Innovation Director at Saatchi & Saatchi and Tom Ollerton Innovation Director at We Are Social.
But what is Computational Creativity?
According to whatis.com ‘Computational Creativity is the application of computer technologies to emulate, study, stimulate and enhance human creativity.’
But applying this to an industry of creatives poses several challenges, the main one being the fear of creativity moving from the control of humans to machines.
So what does this mean for the advertising industry?
I always thought that the biggest challenge facing creatives was the client and their openness to creativity versus the bottom line. So this episode was really interesting in understanding the changing role of creativity and the new challenges technology is posing.
Apparently, one of the problems that currently faces the advertising industry is that creatives don’t understand how to best use the new technologies and tools they are being presented with. Instead, there is a psychological barrier: the worry that these machines can make a better ad than them. There needs to be harmony, or a value exchange between the creatives and the technology in order to get the best and most efficient outcomes.
Is data creativity?
On its own, maybe not. For example, you wouldn’t get the Cadbury’s drumming gorilla as a machine’s interpretation of success for a brand from feeding it customer behavioural data. Feed data into an algorithm and you’ll get a very rational response. Currently, decisions are based on a Creative Director’s whim, her gut feel and the client’s trust. But that’s where the industry needs to see a shift: there needs to be a synergy between the Creative Director and AI.
Data is problem solving, and using specific data sets to align with your creative ideas can be beneficial in making commercial judgement. For example, at Karmarama they are learning to use data effectively in the creative process by using AI to understand what content is performing better for a brand. A creative is still in charge of the concept, but algorithms tell that creative how much more work of a particular theme they need to make.
Budget vs reality
At the moment, most organisations and brands don’t have the budget to create personalised ads (take TV ads for example). Time and resource being the restricting factors. In the future, AI will allow us to create individualised creative responses. However, it is much more conceivable for AI to produce a personalised piece of content based on someone’s interests on social media, than it is for AI to produce an advert that is going to resonate with 20 million people across the world.
And big adverts still have an important role to play, as they have the unique ability to create public conversation and discussion. If all the advertising we were fed became unique and personalised, human interaction around brands would take a nose-dive as we’d have no shared content to talk about.
A historic love for the creative process
Research by Capgemini showed that of 993 companies implementing AI, 74% said it was making their companies more creative. Research by The Drum, who partnered with Sysomos, showed that while marketers feel positively about AI, only 37% are actively investigating implementing it. This suggests that our creative industry is 40% behind other industries looking to use AI as a creative tool.
It also suggests that the adoption curve for computational creativity in the less creative brands and companies will be higher, because they are less emotionally attached to the creative process.
Although in turn this could free up creatives to spend more time on brands in their portfolio that value their brand message over bottom line sales, i.e the work the creatives would rather be doing!
Good examples of practical Computational Creativity
Script book employ artificial intelligence to analyse screenplays. They provide an objective assessment of a script’s commercial and critical success. Suggesting that that AI makes a better judgment than Hollywood’s Creative Directors and Producers.
A weekly roundup of songs Spotify believe to be suitable for you. Based on your listening history and that of other Spotify fans with similar tastes, it suggests a playlist that continues to improve the more you use their service.
The Bronze Composer enables artists to create and record ‘living’ music that can evolve beyond the studio.
I’ll Be Back is a monthly meetup for people who are interested in the impact of artificial intelligence on the creative process in the ad industry.
Good examples of creative Computational Creativity
Nutella used a computational creative algorithm to design 7 million different jars of the hazelnut flavoured chocolate spread. Will we start to see Creative Director’s signing off algorithms rather than the individual pieces of creative? But who would be responsible if these algorithms changed over time as they learnt inputs and outcomes?
Goldsmiths produced a piece for the Whitney museum in New York last year based on machine learning AI algorithm that learnt Blade Runner the film frame by frame. It then re-remembered that film in a machine vision language.
Paul the robot works for MediaMarktSaturn. He learnt where people buy their products and which ones are most popular to the point that he can now greet the customers and take them to their favourite aisles.
Given all of these different examples, it’s fair to say that sometimes we need to re-frame what we mean by computational creativity. At the moment we see creativity and technology as two quite different things, but we should be looking at creativity as a technology.