ARTYficial Intelligence or just Artificial, can tech be creative?

Olivia Brookhouse    27 September, 2019

When I hear the word art, a few masterpieces spring to mind; Monet’s Water lilies, Picasso’s Guernica and Van Gough’s Starry Night. Whilst many artists names’ do not carry the same grandeur of those above, even a few smudges on a canvas can convey a wealth of culture, history and emotion. Art extends beyond paintings to poetry, plays, novels, music, fashion and film and at the helm of these artistic forms are creators, pioneers, innovators and developers. But are they all human? In this blog, we ask what is takes to be creative and whether this creativity extends to machines.

Painting made by AI technologies using generative adversarial network – “algorithms are able to emulate creativity”

The dictionary defines creativity as the ability to transcend traditional ideas, rules, patterns and relationships to create meaningful new ideas, forms, methods, interpretations. In Daniel Kahneman’s book, Thinking Fast and Slow, he explores the two different ways the brain forms thought between System 1 and System 2. Whilst System 2 is logical and analytical, System 1 is impulsive and automatic, formed by gut feelings and evolutionary adaptive tendencies.

A study of creativity

A study by Jonathan Smallwood proved that the most creative individuals were able to harness both System 1 and System 2 to make creative connections. They were “simultaneously able to live in a dream state while concentrating on the outside world”. Whilst imagination (system 1) was important in generating unique ideas, logical thinking (system 2) was necessary to harness these ideas. Machines have the tools to accurately analyse and make logical connections but whether they are able to simulate traits of system 1, to dream, is questionable.

From composing music to building sculptures, AI systems in art are fed data from 1000’s of examples and then it finds patterns and trends via Machine Learning. Finally the system is capable of replicating and creating similar versions. If we look at creativity simply, as ‘transcending rules and patterns to create interpretations’, is this not what machines are doing? In one of our previous posts we talk about how companies incorporate AI into many creative processes.

From an input, the system generates gaussians, a range of possibilities of where the next pen stroke should go. As more images are inputted, the system gets smarter. https://www.youtube.com/watch?v=VKdsh2UvNqc

AI, the biggest copycat

Whilst what we admire about artists is their uniquely recognisable styles, we value machines more for their copying skills. Any originality is purely coincidental.

It’s easy for AI to come up with something novel just randomly. But it’s very hard to come up with something that is novel and unexpected and useful.

— John Smith, Manager of Multimedia and Vision at IBM Research

Remember that artists take inspiration from external stimuli, events and relationships, like how AI interacts with inputted data. AI may not be a pioneer, creating somehting remarkable, but does this mean they do not demonstrate creativity?

The argument on whether AI can be the next Picasso is dependent on how you view creativity, as a process or a means to create something engaging. Many will view a beautiful piece of art and be content. But to many, creativity is a process, actively choosing to move away from what those have done before. Machines will be truly creative when they decide to start drawing without instruction but isn’t that a scary thought.

AI in Art, for the time being is a tool to help make art quicker, an intelligent stencil. Whether we can teach AI to be creative for itself, to produce something ‘uniquey beautiful’ is yet to be seen. True creativity comes from being surrounded by senses and experiences which we experience as humans. We would have to supply machines with everything we experience as humans; smells, sounds, feelings, connections, to become a true artist.

To stay up to date with LUCA, visit our Webpage, subscribe to LUCA Data Speaks. Follow us on TwitterLinkedIn or YouTube.

Leave a Reply

Your email address will not be published.