Artificial Intelligence takes on Creativity

An AI replica Rembrandt, but is it art?

Artificial intelligence is capable of a lot. It can help us to diagnose disease, fight crime and beat us at poker. However, when it comes to creativity, the power of Artificial Intelligence has been doubted. The reason for this is clear – creativity is inherently human. People, purely because they are people, can comprehend emotion and beauty on a completely different level to unfeeling software. Generally speaking, so far creative workers have had less to fear in the ongoing automation of traditional jobs. . . that is, until now.

A project completed by J. Walter Thompson for banking company ING has revolutionised the way we think about creative works by revealing The Next Rembrandt, showcasing a painting designed and created by AI in the style of the artist. After months of analysing original paintings, the algorithm came up with a portrait of a middle aged, white male wearing seventeenth century get-up in the same style as Rembrandt, 347 years after his death. This isn’t the first example of creative AI, and it won’t be the last – but how else has AI been applied to traditional creativity, and how will it disrupt creative industries?

Artificial artists
What does it mean to be creative? Creativity goes hand in hand with innovation and suggests that the artist, director, musician or writer is coming up with something new. Whilst the Next Rembrandt is hugely impressive, the finished portrait is essentially a multi-layered copy. Ron Augustus, Microsoft’s Director of SMB Markets, would disagree. He directly compared the team’s use of technology and data to the way Rembrandt used his paints and his brushes ‘to create something new’. But is it really new? The algorithm threw together various original Rembrandt paintings and then spat out something in between – not new, in fact, but a median. However, if a human art critic was unable to detect that the painting wasn’t genuine, then the project would be a step towards passing the Turing Test.

So, if AI can emulate a world-renowned painter, what else can it do? Well, apparently it can direct films, as shown last year when an algorithm wrote a mini movie called ‘Sunspring’. This year, the algorithm (that has named itself Benjamin) has directed a new and somewhat bizarre sci-fi short starring David Hasselhoff. AI can now also write perfectly coherent articles, which is more worrying for some than for others. The creative capabilities of artificial intelligence are expanding, but how with this change creative industries and the work they produce?

AI disrupting creative industries
In a number of instances, AI is making it easier for creative workers to realise their visions. We’ve already seen examples from art and film, but the list doesn’t end there. Various record labels including Polydor are using AI in music creation, marketing and discovery. In 2016, Google’s AI wrote its very own, minute-long song. In the billion dollar gaming industry, algorithms are now learning how to make games. The future potential of creative AI is staggering. As well as enhancing the quality of songs and gaming, AI programmes could be applied to universal design. AI and 3D printing already merged to generate the Next Rembrandt – if they can produce the same quality work as a famous artist, what’s to say they couldn’t design buildings? This could potentially lead to more effective and long lasting constructions. Even in the world of journalism, editors are currently using AI to help detect fake news. But, of course, there’s an elephant in the room. If creativity is inherently human, then truly creative artificial intelligence suddenly doesn’t seem very artificial at all. On top of that, what happens to all of the displaced creative workers who were previously thought of as safe from automation?

AI will undoubtedly become an important tool in creative industries, but it simply can’t create in the same way that humans do. AI is clearly creative in a sense, but so far it can only follow instruction from true creatives who devise the projects and concepts. The Rembrandt project, for example, uses deep learning to create a copy. This doesn’t mean that AI won’t ever create original work, but quite frankly, that’s a worrying prospect. An AI that can truly create is an AI that can think for itself, and that isn’t necessarily something we want. At the moment, Artificial Intelligence exists as an aid to creativity. Whilst recent creative projects using machine learning algorithms may not currently fit the criteria for original creativity, they could do in future. When this happens, we’ll be yet another step closer to singularity.

Does your business use AI as a creative aid? Will algorithms eventually create entirely original work? Does truly creative AI signal the beginning of technological singularity? Share your comments and opinions.