Most of us saw the viral images of Pope Francis in a white puffer-coat and a black leather jacket trending online earlier this year. The slickly produced images were almost comical because one could hardly imagine the leader of the Catholic Church donning such trendy merchandise. The images were shared and commented upon numerous times before the internet realised something was awry with the Pope’s hands and his crucifix pendant in one of the pictures. Even the lenses of his glasses looked warped. The images were fake and were created by Mexico based content creator Leonardo Alpuche using the Artificial Intelligence (AI) software Midjourney. Intended as a joke, the images fooled us all.

Since last year, many artworks, photographs and portraits of real and imagined figures with brilliant graphics generated with the aid of AI technology have popped up online. How does AI work to create such realistic images? Can artworks made by artists using AI even be considered art? Do they deserve a space in our art history books and art contests?

While some ideas of AI date as far back as the 1960s, when early computers used guidelines to make technical images, the recent boom in AI imagery is very sophisticated and complex. AI technology often works through algorithms and machine learning frameworks, one being Generative Adversarial Networks (GAN). GAN is first fed with thousands of images, their captions and data ranging from paintings and drawings to photographs and illustrations. The programme is then able to construct an image based on its training.

These AI models are like ‘energy-based models’ from the 1970s. As countless images are fed into the software, the AI sucks up different representations and tries to create a new one. The portrait of a fictionalised man named Edmond de Belamy demonstrates the effectiveness of this mechanism. The work was generated by the French arts collective Obvious in 2018. The figure in the portrait has a chubby face and wears a black coat with a white collar. The blurry ink painting was made after the AI software referenced 15,000 existent images. However, let’s not forget here that AI can only generate images from examples of what humans have already created throughout history.

The rapid increase in AI-generated artistic images is leading to debates around creativity, originality and authenticity

Though AI models can provide us all with a good laugh and a creative perspective, they can also set dangerous precedents around the existing concerns of authenticity and validity. As exhibited by the image of Pope Francis, AI technology has thus far contributed to the already existing pool of fake news and misinformation.

Consumers of art have decried the inclusion of AI-generated images in art contests. As scholars, we question if AI-led imagery created by human beings may even be permitted in competitions where creative practitioners mould technology and tools to create artworks, as opposed to feeding thousands of images to a software that makes something out of them. For instance, Jason M. Allen’s AI-generated work Théâtre D’opéra Spatial (2022) won in the emerging digital artist category at the Colorado State Fair’s annual art competition. How are principles of impartiality to be considered in such a case? People wonder if (in Allen’s case or any other) it is the AI-software that wins or the individual.

As you are lulled into triple-checking the authenticity of your Twitter feed (while questioning your sanity too), AI has started to garner some unprecedented institutional representation. The Mauritshuis Museum in The Hague displayed an AI-generated image of Dutch painter Johannes Vermeer’s painting The Girl with the Pearl Earring (1665), made by Berlin-based creator Julian van Dieken using Midjourney. Titled A Girl with Glowing Earrings, the work is a fan-made representation of the globally beloved Dutch historical masterpiece. Art markets have also responded to AI as the Belamy portrait mentioned earlier fetched a sum of $432,000 (much to the dismay of many) in an auction led by Christie’s.

Fake reality is regrettably turning into a new reality and the world ought to remain wary. As AI continues to infiltrate our visual culture, we must be ready to anticipate new philosophical questions about authenticity and fairness around non-human agents.

The writer is a Fulbright scholar, art historian, art and book critic, and an industrial designer.

She can be reached on Instagram @pressedpulpandink and Twitter @nageen_shaikh

Published in Dawn, EOS, June 4th, 2023