At a concert organised by the University of Oregon, artificial intelligence named EMI managed to deceive its audience by passing off its musical compositions as pieces by Johann Sebastian Bach himself. In New York, a computer programme wrote a novel from start to finish inspired by Jack Kerouac’s legendary On the Road. In Paris, Edmond de Belamy, a portrait painted by an artificial neural network, was auctioned for thousands of euros.

As new and groundbreaking as all this may seem, it is actually part of the prehistory of artificial intelligence (AI). Both the auction and the publication of the novel took place five years ago, in 2018, but algorithms like that of EMI have been in development since the 1980s. The obsession with creating a machine in the image of the human brain, an intelligence capable of calculating, reasoning or even creating artistic works goes back a long way. In recent years, however, steps have been taken at such a speed that the line of what we thought was inconceivable has shifted.

In 2022, AI programmes such as Dall-E2, Midjourney or Stable Diffusion surprised us with their ability to create unprecedented, instantaneous and astonishingly realistic images from a simple text description: from cowboys on the moon to cubist versions of The Girl with a Pearl Earring or recreations of the French Revolution as seen through a fisheye lens. Others, like GPT-3, demonstrated their ability to write complex texts and original poems within seconds.

Just as EMI was able to imitate Bach, there are algorithms that can compose their own melodies, illustrate comic books or write scripts for short films.

Such algorithms are capable of defeating human ingenuity in an artistic contest, as one of their mechanical predecessors did against an unbeatable chess player, Kasparov.

The creative AI revolution has sparked a wave of fascination, but also fear, and not without reason: if 56 per cent of today’s jobs are already threatened by automation, could artists, designers and creative professionals be next in line? And to what extent is artificial creativity capable of making human creativity obsolete?

Where does artificial creativity come from?

Like the human brain, artificial intelligence-based programmes have their own networks of neurons – artificial neurons. Huge amounts of data are stored in them, which are then analysed, classified and recombined. This is how ideas are hatched.

“The big leap in artificial intelligence came a few years ago with the development of deep learning. The support was acquired to train deep networks within very short timeframes so that they could not only analyse data but also learn from their own experience,” Karina Gibert, director of the IDEAI-UPC artificial intelligence research centre and one of Spain’s leading specialists in this field, tells Equal Times. “We call it creative intelligence, because it is able to take separate elements acquired through learning and combine them to generate something new.”

One example is the Edmond de Belamy portrait, created by AI after analysing 15,000 portraits painted between the 14th and 20th centuries. Another is The Next Rembrandt project, which, after studying 300 works by the Dutch painter, is capable of painting new works recreating his style. The same procedure is followed by programmes such as Dall-E – a fusion of the name of Pixar’s WALL-E robot and Salvador Dalí – which create images from text. The user writes what they want to see and the AI searches its database and designs several versions within a very short space of time.

AI’s strongest critics accuse it of copying, of practising a sophisticated form of plagiarism. But what it does is not very different, in fact, from the unconscious process developed by the human brain. As Margaret Boden, a leading figure in artificial intelligence, explains, new thoughts that originate in the mind are not likely to be completely new but rooted in representations that already exist . The fact that machines have learned to replicate part of our creative process does not mean that they can emulate it in its entirety.

“There is an element of creativity that cannot be automated,” María Luisa Sanz de Acedo, professor of psychology specialising in thinking skills and creativity, tells Equal Times.

“Some call it subjectivity, others life energy, spirituality. AI can simulate some human behaviour, but that complex inner world is beyond its reach.”

Pablo Sanguinetti, author and researcher on artificial creativity, agrees. “AI doesn’t imagine, feel, have consciousness or understand the world,” he explains. “These systems do one thing extraordinarily well: processing large amounts of information and extracting patterns from that apparent chaos; but what defines creativity with a capital ‘C’ is the human capacity to break free from those patterns, reject them or re-signify them. I’m not sure machines are capable of this yet.”

Tool or threat?

Cristina de Propios is a Spanish artist and designer. She has been focused on innovation since the very start of her career, combining art and robotics. It was only a matter of time before she moved towards artificial intelligence. “When the Belamy painting was sold, it really grabbed my attention. I asked myself ‘how can a machine make art’, but before criticising it I decided to give it a try.”

Her latest project consists of “living canvases” made up of cellular automata that, thanks to deep learning, create different compositions and movements each time. “This technology is very interesting in terms of the visual possibilities it offers and its surprising results. It’s a kind of augmented creativity,” says the artist, who insists on highlighting the importance of the human factor. “AI doesn’t understand what is beautiful or harmonious. It is like a small child that you teach based on what you know.”

Like her, a growing number of creators are using artificial intelligence as creative tools. “The results the machines generate by combining data can act as a stimulus in the creator’s working memory,” says creativity expert María Luisa Sanz de Acedo. In other words, AI can provide resources to avoid having to start from a blank canvas.

“Just as some people go to the sea and others read books, AI can also be a source of inspiration,” explains computer scientist Karina Gibert.

In sectors such as design, however, they do not necessarily see it as favourably. The fear is that the possibilities offered by these advances could be exploited to cut costs, demand more work in less time, to make their sector even more precarious or to automate the simplest tasks. “I think some jobs are definitely at risk,” says Abel Guzmán, a graphic designer. “As with other technological advances, new professions will emerge. There will be people who specialise, who learn to understand how AI thinks, and how to direct it. Personally, in my case, I think I only have two options: to train or be left behind.”

Reskilling, he points out, will not necessarily be easy, especially in a sector that is evolving so fast and where it is difficult to compete with new generations. This is the reason behind the recent protests by creative professionals, such as the manifestos posted on social media, where some have created their own seal to identify their works as AI-free. “I can understand that,” says Guzmán. “Many of these people are older or have a very purist view of the profession. They’ve seen the dangers on the horizon. It’s true, though, that we offer something artificial intelligence cannot. AI produces results but it doesn’t talk to the client, it doesn’t know the emotional charge of a project. If you are looking for a mechanical job, you can ask AI to do it, but if you want a more human, more empathetic job, you should know that you won’t pay the same as to someone who only has a computer farm.”

AI and new myths

The debate over AI’s impact on employment is not the only dilemma presented by creative artificial intelligence tools. The matter of who owns the copyright to these works, for example, remains pending. Is it the AI itself, the person who gives it the instructions, the creators of the works from which it draws its inspiration, or the computer scientist who developed it?

“The rise of a new creative technique not only introduces new forms of production, it also changes the very idea of what art is,” says researcher Pablo Sanguinetti.

“You just have to think about photography, cinema, the printing press. There is always a degree of anxiety at such times: the fear that the technique will replace the artist.” It is the same story already told by science fiction – from Blade Runner to Frankenstein – hundreds of times.

“The fear that people used to have of fate, of gods, of providence, has shifted to a fear of ourselves, of our world of objects. The myths surrounding artificial intelligence are like those that used to surround saints, such as the power to perform miracles,” Fernando Broncano, professor of philosophy at the Carlos III University in Madrid, tells Equal Times.

For the professor, there is no sense in fearing machines, or in thinking that we are fundamentally very different from them. “We are already part machine. The human species is as much a product of its biology as of its technical environment, its tools, its medicines.”

“Life is the greatest example of creativity that exists,” the philosopher recalls. “For example, a storm can produce a change in the environment, it can generate something new, but it is the human being – society – that gives it value, that decides what to do with it. The focus of our thinking should not therefore be on what differentiates us from machines but what we want to do with them.”