Liminality between original creation and plagiarism: the use of AI to generate art.

The services provided by artificial intelligence to form images that, albeit mesmerising everyone, carries out a process that remains unknown to many; we know image recognition is the main step undergone for the production of A.I. generated art, but for the vast majority of the population, that’s about it. This lack of awareness commences with the void of knowledge in regards to the beginning of artificial intelligence. Although many people believe that it began with a website in the devices we hold in our hands, the use of A.I. to generate art dates back to the 1950s and 1960s, when computers were used to create colourful patterns. Three decades into the future, this tool would have experienced notable developments, and individuals would begin to make use of this asset in different sectors within the arts. Artificial intelligence was being used to create innovative pieces in literature and in the music industry. Furthermore, since the most recent development of A.I. features, and the presence of NFTs in the digital creative industries, A.I.-generated art has become a common option as a purchasable product in this new format, meaning that anyone can create images with this powerful tool and gain revenue from them. Moreover, the rightful use of these images for commercial purposes derive from the fact that they are granted by the creating softwares as copyright-free media, which initiates the steps into the ethical debate of this tool to generate ‘new’ art. One of the most intricate aspects of this tool in the modern day is the deep learning methods it follows, making use of immense databases and visuals and media recognition in a way that mimics as loyally as possible the human form of processing. This is achieved through the use of deep neural networks, a process in which A.I. gains different types of data to achieve said repetitive behavioural traits. Additionally, the extent to which text-to-image prompts work is impressive to all, and the images generated sometimes push the limits of reality in ways in which we would normally deem impossible from a non-creative to produce, and much less a machine. Processes undergone in prior decades showed a conceptualisation process which, mentioned in subjectivity, is lacking in the generation process we follow today to create A.I. generated art. Presently, efficient machine learning algorithms are used. For example, a simple prompt such as ‘Portrait of Salvador Dalí, painted in the style of Pablo Picasso’ suffices to compose an image with sufficient level of detail, as well as understanding of the provided inputs. In th4e beginnings of the use of this tool to produce art, there was a concept generated and developed behind the final product. Now, we can get the previously mentioned image of Dalí in two minutes or less- the same amount of time it took me to think of that concept for the piece. Illustrator Rob Biddullph wrote for The Guardian that A.I.-generated art is “is the exact opposite of what I believe art to be.” Biddullph argues that AI generated art lacks the emotion of the creator, which in some cases may not be true, but simply by looking at trends of sharing art generated by artificial intelligence on the internet, I believe we might be able to say that most generated pieces are created for the user’s entertainment. In other words, mostly, this art does not have years, months, or even days of experience carried by an artist. Oftentimes, the effort in constructing a concept for a piece of art is valued to a much bigger extent than the final product. As someone that values art deeply, especially all the process it takes to create a single piece of work, I believe it is important to state that the difference between the conceptualisation and production time between generating AI images and art made by humans is abysmal. Although some A.I. pieces do have deep and intertwined concepts behind them, many others are devised in less than five minutes. With the generation of art through artificial intelligence, new pieces can be generated with minutes in between each creation, which also poses the question of whether this will lead to the devaluation of art itself and all the work that goes behind it. To put it in simpler terms, creation of art using artificial intelligence can be overall considered superficial, as it generally lacks profound meaning, technical expertise and emotive reasoning. Although some parties counteract and explain that original art pieces are merely used as training data, A.I. tools make use of billions of visuals, belonging to other artists, to form new images and furthermore, share it free of copyright, making us truly question the validity of this art and the use or referencing of pieces of independent artists without granting adequate credit. As with everything, the complexity in the way individuals experience and view ethics within society is so convoluted that determining the practise of this tool as utterly good or bad is simply not possible. Nevertheless, the development of this tool and the extent to which A.I. generated art is beginning to occupy spaces previously given for artists to showcase their work already gives us a hint of the path to which we are being led. Although we cannot measure precisely the benefit or detriment of A.I. to create art, the similitude in style between all of them is simply the upper layer of repetition of image recognition patterns with no conscience or emotion behind them, I believe the question of whether it is truly ethical or not will be questioned more seriously when these pieces take up space given to artists in order to share their work and artificial intelligence not only serves as an additional tool, but as a method to replace the work artists create.