Until recently, artificial intelligence (AI) was a distant, dreamy concept that we dusted off annually during Christmas reruns of Bladerunner. However, in recent years, it’s gradually crept into our lives in very tangible ways. From the Google Home device in the corner of your kitchen to automated supermarket tills, we’re beginning to see AI being put to exciting uses in disciplines that require high-level thinking. And the creative industry has been no exception.
However, as creative bots start to become more widespread, many are left wondering whether AI can ever have the creative freedom or intelligence needed for unique and valuable innovation. In other words: how far really are we from developing a strong AI algorithm that can independently identify and bring to life unique creative concepts? Firstly, let’s take a look at the range of ‘creative’ AI bots currently available to see how they measure up.
The artist: Molly at TheGrid.io
What does she do? Molly is the AI web designer powering a website building platform called ‘The Grid’. Molly claims that she can identify and build an appealing design for a website based on her analysis of patterns within the content.
Under the hood: With generic template websites becoming increasingly prevalent, it’s easily conceivable that a bot could fill the role of churning out another mediocre website template design. It’s less likely, however, that she’d be able to capture the hearts, minds, and purchase intent of the site’s visitors as easily.
The artist: Nutella’s algorithm
What does it do? Ferrero, the producer of Nutella, employed an algorithm to design a limited edition collection of seven million unique packaging designs sold on its jars throughout Italy. Not even Picasso could have rivalled that kind of creative output.
Under the hood: Ferrero’s agency supplied the algorithm’s software with a database of tasteful patterns and colours that they felt would fit the brand’s spread. Would the delicious jars have looked so appealing without the discerning creative choices of the minds behind the software?
The artist: Portrait
What does it do? Portrait is an app developed by the team at Img.ly, and creates beautifully designed self-portrait images using supervised deep-learning and the visual power of its software development kit (SDK).
Under the hood: Portrait has definitely nailed the ability to do some clever automated image masking, clipping, and colour adjustments – but it’s safe to say that the bulk of the serious creative work here has been delivered by its very human, creative director Tommi.
Our fears are short-lived
Thankfully, the above highlights that any fears for our creative jobs today are short-lived. Customers at The Grid quickly discovered that Molly couldn’t sufficiently solve many real UX challenges or deliver innovative creative work, and without the human designer’s tastefully selected patterns, Nutella’s jars would have maintained their standard branding.
Experts say that we’ve only just begun to scratch the surface of what’s possible for AI and creativity. Creativity is a challenging subject, which often evades definition, but can be summarised as the act of turning new and imaginative ideas into reality. Creativity involves two very distinct processes: thinking and producing. We can see from the above examples that we’re beginning to develop the skills required to master the production aspects of creativity. But the fact remains that our current technologies are still limited to weak AI.
Weak AI vs strong AI: What’s the difference?
Weak AI: Weak artificial intelligence is a form of AI specifically designed to focus on a narrow task and to seemingly do so very intelligently. Today’s bots often excel at specific creative tasks, but only within very rigid confines.
Strong AI: By contrast, strong AI is capable of any cognitive functions that a human has and is, in essence, no different from a real human mind. Without strong AI, bots will never be able to attain the freedom and higher level of thought needed to recognise and shape real creative concepts.
How can we bridge the gap?
It’s clear that there’s a big gap to be addressed before any AI will encroach on creative thinking. For example, we’ve long asked whether AI can truly be taught what’s beautiful? Many would argue that beauty is a subjective concept, and that knowing what’s ‘beautiful’ is an ultimately impossible task for an AI bot. However, with the evergrowing volumes of readily available user generated data, this possibility is becoming increasingly real. If an AI bot is given enough training data, it can learn rapidly what fits within the commonly accepted range of ‘beautiful’. Amazon has already started harnessing this ability and built an AI tool that can intelligently detect and replicate fashion trends.
More challenging still, can an AI be taught to identify a novel and useful concept? Products and teams such as Adobe Sensei are focusing on gathering and analysing data outputs from designers to learn how to mimic and optimise on our creative processes. We’re quickly gaining ground in finding ways to harness bots to reduce the manual labour involved in creative production. But an AI bot with a higher level of thinking that can understand complex, diverse briefs and identify valuable creative solutions is still only a fictional dream.
So, don’t fret, the research shows that, at least for now, creatives are safe.
Catherine Pearson, Lead designer at Greenlight.