AI, Machine Learning and the Creative Industry
This week Jon is looking into AI and Machine Learning and specifically how these changing technologies could affect the creative industry.
AI has seen massive leaps forward in the last 12 months. Wired Magazine recently reported that machine learning is tracking websites, news reports, and social media posts for signs of Coronavirus symptoms. This follows news reports that AI can identify cancer more effectively than trained and skilled pathologists in a groundbreaking and big development. But will we get to a point where AI can do everything?
I’ve recently been enjoying Malcolm Gladwell's new book, 'Talking to Strangers'. In it, he writes of the work of the economist Sendhil Mullainathan. In Mullainathan's experiments, he found that removing the subjective element of how one ‘feels’ about a defendant in court enabled the AI to be much much more accurate at putting likely repeat offenders in prison than the judge. It wasn’t even close. With a fraction of information to the judge, the program made a better job at making bail decisions. In this case, feeling and intuition, which is a massive part of the creative process, was removed making it more vastly more successful.
After reading about how by removing some of this intuition and feeling could make a subjective process better it led me to wonder, could AI or machine learning learn creativity? Creativity is all about feeling and intuition. From birth, we are innately creative beings. Ken Robinson’s now-famous early TED talk talks so articulately about creativity in children and how today’s education system educates us out of this natural creativity. My experience of visiting all those art exhibitions, seeing all those films and theatre productions, listening to all that music is utilised every single day undertaking complex narrative creative endeavours. However, with the automation of all human jobs predicted in the next 120 years, could a machine steal my job?!
On a mechanical level, RTVE, a national Spanish TV channel is testing an AI programme that can pull archive clips into an edit timeline within seconds. So can the AI learn good editing, tell the story and create an emotional well-crafted piece with well-paced cuts to visual cues and to music? Maybe not now but it will only get more and more intelligent. This is something that I can see the machines doing quite well. Taking this one step further, can a computer direct a movie? Can a computer understand and manage all of the creative components of the creation of film and sell that concept to raise vast amounts of money to make this complex production. Maybe the algorithm in time will hit all the right points to release funds from a bank...who knows what the future will bring...but directing a movie such as the Oscar-winning Parasite with all its cultural nuances, twists and turns and social satire? I don’t think so.
What about scriptwriting? The TV industry has been revolutionised in the last 10 years through the advent of Netflix, Prime, Hulu, Disney+, Apple TV etc. It has changed the way most of the world’s population engage with creativity through the making of television. This is a vast industry constantly charged with coming up with new and innovative content. For the first time, we are seeing films and TV shows being made based on vast quantities of data from streaming services such as Netflix and Prime. We used to have some basic data from organisations such as IPSOS MORI, but now these streaming services are producing very detailed information on our viewing habits and what we like and don’t like. They can even use this information to change the thumbnail on programmes and films to make them more appealing to us depending on previous viewing habits. There is no doubt that shortly TV scripts will be written, songs will be recorded, animations will be made without human intervention. But will the audience connect with them at a gut level, pulling them by their hearts and minds along an emotional journey that will inform and move them?
Whilst doing research for this article I found my favourite take on AI vs creativity thanks to Nick Cave. Cave opened his process up to his fans through his site the Red Hand Files where he answered the question ‘Considering human imagination the last piece of wilderness, do you think AI will ever be able to write a good song?’. The short answer is yes but it won’t be comparable to Prince, Iggy Pop or Beethoven. He writes, ‘What we are listening to is human limitation and the audacity to transcend it’. I choose to believe that the uniqueness of the individual experience within the great practitioners of the creative endeavour is not transferable through machine learning. The experience of the pain and anguish from a performance by Nina Simone can only be conjured through personal experience. Nina Simone was a one-off and no machine could create that.
Even the vast computing power that enables the world of deep fakes and AI-created content will never be able to equal or better the dark and wonderful minds of filmmakers such as Peter Greenaway, David Lynch and David Cronenberg.
We are already seeing great strides of utilising data and technology in the highly commercial and very creative world of advertising. ROI has been king in advertising for a number of years but we are far from the golden age of advertising where the ads were as creative as the shows they inhabited. At the moment there seems to be too much focus on technology and AI. This raises the risk that advertising style and creativity takes a big dip and becomes less effective. The outcome is fact-driven and product focussed ads that move metrics over the short term rather than creative and emotional ads that move audiences and have lasting impact.
I don’t reject machine learning and welcome the incredible ability that AI allows us to understand how audiences appreciate and receive creative work. We will move forward hand in hand with this new technology leaving the mundane to the machines and the deep creativity to the individual unique minds of the creative innovators.