Does AI Have Creative Potential?

Scientists developing artificial intelligence (AI) are quick to point out that teaching computers to be creative is a different methodology to how the human mind can be inventive. On the flip side, German artist, Mario Klingemann says “Humans are not original.” We typically find inspiration from what has been created in the past.

The question on everybody’s lips in the tech space is whether AI can also be trained to be creative without guidance from humans. Opinions differ. Machine learning is causing a stir in some quarters. Perhaps persuaded by science-fiction films such as Terminator and Ex Machina, some commentators are concerned the rise of the machines will overpower humans in the future.

Captain’s of industry such as Rob High, Vice President and Chief Technology Officer for IBM Watson on the other hand say the goal for AI is not to recreate the human mind but to teach computers to inspire humans. There is still some debate as to whether computers may develop the ability to be self-learning, however. In 2017, the AlphaZero AI created by Google’s DeepMind taught itself the rules of chess in four hours and consequently beat the world’s Grandmaster.

With a basic understanding of the rules, AlphaZero was able to work out the game by consistently playing and reinforcing knowledge. The difference between rules and creativity, however, is that the former has a finite permutations. Having said that, AI has even entered the art space. A piece of AI-generated art sold at Christie’s auction house in New York for as much as $432,500 – albeit reports indicate a significant amount of human interaction was also involved.

But then there’s the world’s first robot artist Ai-Da churning out paintings at Oxford University. Although at the moment, she is only able to analyse an image in front of her, engineers say her goal is to be creative.

Can Deep Learning Be Creative? 

Empirical evidence shows us that deep thinking leads to creativity. Although neuroscience doesn’t know quite how the neural networks conjure creative thoughts in the mind’s of man, innovation nevertheless occurs all the time. In general, however, as Klingemann points out, most humans mimic what they already know because the human brain is generative. We tend to repeat what we read and hear in parrot-fashion.

Deep learning technology is also a generative model. When you feed AI data, complex mathematical systems learn how to develop permutations. If you keep tweaking with parameters, the number of potentials continue to increase. Whereas humans are quite apt at making sophisticated decisions by taking into account convoluted considerations, when the number of possible solutions increases, AI is far better equipped to make decisions efficiently and consistently.

There is, therefore, the potential for AI is to unlock the creative potential in humans. It can give us more options, insights and awareness to help encourage higher-level thinking and strategic decision making. If we define creativity as the ability to come up with new ideas then, yes, AI can be trained to be creative. But how accurate will a computer be? There is no means for a computer to understand human emotions and feelings – which we know play a huge part in a creative process and in the way an audience reacts.

Stimulating AI Emotions

As audiences (and consumers) continue to have a shift in perspective and demands, artists and businesses are challenged with the hurdle of engagement. Understanding human emotions is not an area with which AI can help us.

Or can it?

Disney has been experimenting with AI in cinemas since 2017 to register the reactions of viewers. Tech giants and government border controls have been using facial recognition software which has eye-tracking and facial expression features which analyses mood.

Researchers have already used machine learning to identify emotional arcs in stories. McKinsey report that AI has the potential to enhance plot lines by suggesting emotional pulls a writer could use to enhance storytelling. However, computers still lack the creative intelligence to piece an entire script together.

There’s little doubt that AI can provide a creative arm in a diverse number of industries, but the strategic combination of man and machine is still required for validation. What AI can be used for is to prompt humans to be more creative by revealing complex links that are not always obvious. It can sometimes take 200 years for someone to identify the next piece of the jigsaw.

Whilst we shouldn’t expect to see any AI-generated Pixar movie hits rattled out any time soon, the capacity for computers to provide a spark of inspiration for humans may help key decision-makers, corporations and scientists to develop technologies that improve the world in key areas such as energy, medicine, and finance.

Sliced Bread’s Closing Comments

We are a long way off from AI successfully developing content that guides audience emotions to a positive result, there are way too many variables and factors that would need to be considered in the machine learning process. Film particularly, has a subjective view, whilst popular films appeal to wide audiences, you can be sure that its popularity is down to a multitude of factors, including individual states, not just collective – i.e. it reminds them of a particular experience in their life, past or present. It is unlikely that AI will be able to steer human emotions, whilst it is fair to say it won’t be generic, tear-jerkers, comedies, horrors, will all require craft, and audience feedback, all of which AI would need to take into consideration in order to sustain audience attention. Also, how big a deal would it be for the studios if an AI engineered film was flop at the box office? Probably huge.

“When we engage in creative works, we take full advantage of all the data that we have processed in the past. Computers are very bad at doing that, especially if we want computers to write each word, each sentence on their own. There are too many combinations of words, sentences, and thus the problem becomes too complex for computers to solve, at least for today’s AI.” Reference link.

Quite often creators want to elicit an emotion, a message, or a particular takeaway from their work, and therefore for creative work to have some kind of success, AI would need to have an intention, and be fed background experiences to be able to emote that. Because AI needs that ‘feeding’, for now, a 100% AI experience for storytelling, animation, film, theatre, is a while off yet.

If you would like to know more about how AI can support your business objectives, drop us an email at info@sbanimation.com, or give us a call on +44 (0)207 148 0526