First, machines took over manual labour. Spinning jennies, steam engines and washing machines were controversial technologies of their time, but looking back, their development doesn’t seem that surprising - after all, humans have been using tools to augment their physical abilities since the dawn of our species.
Then we succeeded in creating machines that could perform cognitive tasks, tasks that we so easily forget were originally performed by people (as the award-winning film Hidden Figures reminds us). This was more uncomfortable; we’ve always identified our intelligence as what separates us from other beasts.
But 2018 will be the year that machines break ground in the final, most taboo frontier: creativity. Creative intelligence has always been characterised as uniquely, even sacredly human - are we ready to share that with machines?
The list of traditionally human abilities that machines can perform well is continually expanding, for example, artificial intelligence (AI) now rivals dermatologists in identifying skin cancer. The success of deep learning’s performance in replicating human cognitive skills (or even outperforming them) has even surprised those working in the field. While some applications are clearly problematic, we have become more comfortable with machines playing a greater role in our lives. The ‘Godfather of deep learning’, Geoffrey Hinton, thinks this is only a hint of what’s to come, asserting that: “Any old classification [recognising patterns] problem where you have a lot of data, [is] going to be solved by deep learning”.
But surely creativity is more complex than "any old classification problem"? Alan Turing was fascinated with the idea of a machine that could write a sonnet (although he added the caveat that “a sonnet written by a machine will be better appreciated by another machine”). As far back as 2011, an AI-written poem was accepted by a literary journal. However, it’s questionable how much this is a true win for computer-authored poetry, or to what extent it simply reflects how well-suited poetry is to this test, given its underlying “ambiguous and bizarre” algorithmic nature. There’s an important distinction to make between programs which can mimic the creation of existing artforms, and programs which genuinely push the boundaries of creativity itself.
So, why will 2018 be any different? We are not just seeing the development of programs that can replicate what already exists, AI is beginning to be used to take us to places we’ve never been. Generative design is an AI-aided design process where, for example, a suspension bridge designer inputs goals and parameters (such as size, materials, load-bearing capacity) into software which then explores all possible design solutions, generating thousands of alternatives in seconds. Many of these are solutions which no human designer could have come up with - the AI’s ‘evolutionary approach’ takes creativity beyond simply mimicking where human designers have gone before. However, Jeff Kowalski, CEO of California company Autodesk (which pioneers this kind of software through projects like Dreamcatcher), asserts that “these technologies are not a threat, they're more like superpowers” for designers. Although the AI generates the possible designs, the human designer plays a crucial role in the creative process, by both curating the design space parameters and selecting preferred outputs. Sound artist Laetitia Sonami explains that using machine learning to make music is less a matter of performing with “a well-trained circus animal”, and more “like riding a bull” - directing the work in unpredictable ways through an element of creative agency.
It is this kind of co-production between AI and artist that will produce Turner Prize-winning work. Maybe the duo will have a creative process that, as with Dreamcatcher, uses a generative model, like artist Mario Klingemann, who builds image-generating software that creates haunting, distorted portraits, from which Klingemann then selects the best for show. Or perhaps the AI’s input will be more insidiously controlling but less superficially explicit, such as in Erica Scourti’s video work Body Scan, in which the artist documents her own body with her phone camera and uses an AI-powered image recognition app to identify related images.
As AI curator Luba Elliott notes, outputs from machine creativity are ‘rooted in a machine perspective that can be vastly different from the human one.’ The exciting aspect of art made with AI is that it is more than a tool to help you create what’s already possible - it pushes beyond the boundaries of human imagination to open up new possibilities of what can be created. In 2017, the Google Brain project Magenta launched NSynth, an AI synthesizer which invents thousands of new instruments no human has ever heard before by combining different ones; you can now play the bassoharp (bassoon + harp), the piccolar (piccolo + sitar), or any other permutation you want. NSynth is one of many creative AI tools Magenta is developing for artists, and as project lead Douglas Eck explains: “We don’t know what artists and musicians will do with these new tools, but we’re excited to find out. Look at the history of creative tools. Daguerre and later Eastman didn’t imagine what Annie Leibovitz or Richard Avedon would accomplish in photography.”
What will artists accomplish with AI? Winning the Turner Prize will only be the tip of the iceberg. Although an award for British artists, the Turner Prize has global prominence and a track record of trailblazing radical approaches to creative process, authorship and ownership - all highly pertinent to an AI and artist duo. Previous winners have often attracted scepticism about their works’ (or their own) value and legitimacy, yet equally frequently, those winners go on to become cultural treasures such as Grayson Perry or Oscar-winning director Steve McQueen. In a period of uncertainty about the UK’s place in the world, what better time to promote two of the industries it currently leads in - art and artificial intelligence? The Government has already made a commitment to advancing AI research. Creative AI could be where the UK comes top of the leaderboard.
Illustration: Peter Grundy