We live in an age where artificial intelligence and machine learning is commonplace. Computers are not only thinking for themselves, they are building themselves. In the arts, in computer science and even in our day to day lives, AI is no longer just a form of human-written algorithmic software – programmers are increasingly finding it difficult to fully understand the machines they once helped create. As developers face this opacity, creativity is becoming a vital distinguishing feature between man and machine. The Turing Test, named after the renowned mathemetician Alan Turing, is used to help make this distinction, and requires that the machine fool a human into thinking it too is human when interacting in a conversation. To date, this has been the most common way to draw the line between machine intelligence and human intelligence.

Computer technologists at the Georgia Institute of Technology suggest that a more compelling test would be for AI systems to produce a work of art that an outside expert would judge to be manmade. As early as 2001 a british computer scientist Simon Colton created a programme called ‘The Painting Fool’ that would paint portraits of sitters according to its mood after having read the morning papers. Some days it would become so depressed by current affairs it would refuse to paint at all. Most impressively, the system began to learn things it had never been programmed to know: ‘The Painting Fool’ began to associate less pattern and colour with worse moods, using pencil on days when it felt more depressed. The machine was able to reach a certain level of self-awareness, assessing its own work and re-evalutating the ways in which it created it.

The intersection of art and technology is no new phenomena. In the 1960s Desmond Paul Henry, a philosopher at Manchester university pioneered a series of drawing machines from modified bombsight analogue computers used during World War II. In 1973 a project entitled AARON was initiated by Harold Cohen, a programme which originally created largely abstract paintings and drawings similar to Henry’s machine, but has since evolved to produce more figurative work and large-scale murals. However, AARON cannot learn new functions and a programmer must hand code every movement made. AARON’s creator Harold Cohen has consistently avoided using the word ‘creative’ when discussing the computer program, but nonetheless wishes to compel its audience to consider whether or not the work produced via his software constitutes art, and why.

Our anxieties and curiosities about the possibilities of AI and its implications for the human race are more pressing than ever before as the evolution of hyper intelligent systems becomes more advanced and the vision that computers have becomes closer to our own. The work of Ian Cheng confronts this issue and is informed by the artist’s personal experiences and curiosity concerning artificial  intelligence. Cheng previously worked for a special effects department in Hollywood until the pressures of creating exact replications of reality drove him to leave. His work is a forceful reaction against this: the artwork, ewCloud (Fatima & Zahra) (2013) is composed of a series of algorithms that toss together a selection of strange and diverse virtual elements. An office chair,  a dinosaur skull and a heron are thrown into the mix, colliding and and crashing together. Cheng’s computerised vision (wearing an Oculus Rift VR headset is the most common way to experience the work) generates an alternate universe where our grip on reality has been lost and instead chaos reigns. In a space where algorithms are set free, dimensions shift and the laws of physics are defunct. As we consider how to process this, Cheng tests and reveals the limits of human control.


Reena Spaulings is concerned with the concept of authorship in a technologised world where the hand of the painter/creator is being progressively erased. Later Seascapes (2015) is created by aa iRobot Scooba 450, usually used to mop floors, Spaulings lets this computational mechanism operate like an artist, making sweeping, gestural brushstrokes across the canvas. Thus, Spaulings allows the machine’s movements to dictate the aesthetics of her work, transferring authorial control. Creative agency is transferred from the traditional author (the artist) to the machine and so the medium becomes the protagonist. This example reveals the synthesis between systems that use ai and traditional abstract art.

Both Spaulings and Cheng expose the limits of the distinctions that seperate the creator from their creation. AI systems have come a long way since Desmond Paul Henry’s first experiments and Harold Cohen’s computer compositions. ‘The Painting Fool’ proved to the world that computers could think and feel for themselves and create art depedant on human emotions. Artists are now able to imbue machines with creative responsibilities and so to shatter hierarchies of technological control. AI presents a real problem to the art world: it can be both a tool and equally it has the potential to be a subversive force. Will a computer’s artistic vision eventually become more intriguing than that of an artist? The future of creativity is uncertain, but we are likely to see many more artists producing work in this anxious and excited state of expectation when faced with artificial  intelligence.

Works by Jeff Elrod and Reena Spaulings