Creativity is one of the most elusive human concepts to define. What is it? Where does it come from? Is it purely a natural gift requiring little practice or is it the result of years of finely honed talent? Joining these much-discussed questions is this – “What happens when AI starts to create?” The answer, much like the question of creativity’s origin, will remain a hot topic of discourse for many years.
Recent trends in the art industry point to the divisive nature of the subject of Creative AI. The Colorado State Fair recently awarded artist Jason Allen first place in their digital arts competition for his work “Théâtre D’opéra Spatial”. As later revealed by the winner, his entry was initially generated using Midjourney – an AI system that creates realistic images from text prompts. The reaction to his stunning revelation was polarizing. While some hailed his innovative approach, some others regarded the use of Creative AI as an intrusion into the art world and a sign of the beginning of the end for human creators.
Midjourney is among an exciting group of emerging AI tools like DALL-E 2 and Stable Diffusion that utilize Artificial Intelligence techniques to birth realistic images from simple text instructions. These tools have been trained on millions (or billions) of images pulled from the web.
Considering the simplicity of the process for the average user, it is easy to understand why the creative industry might kick against labeling the technology as a creative process or even acknowledge the output as an actual work of art.
For a process not well understood among humans, how can an artificially intelligent system be said to demonstrate it?
Writing and music are other creative endeavors witnessing AI’s introduction. In 2019, the album “Proto” was released. It contained the first recorded debut of AI on a pop-music album. Open AI’s deep learning AI writing model (GPT-3) creates entire paragraphs that are indistinguishable from human writing. As shown by these examples. AI’s involvement in the creative industry is widespread thus perfectly setting up the question – “Is human art dead?”
The Role of AI in the Art Industry
All through its evolution as a scientific discipline and transformative technological tool, AI has faced stiff opposition from parties that consider its introduction a direct replacement of human skill. However, as history has shown, AI actually augments labor by automating mundane tasks and freeing up more time for people to indulge in more productive and creative pursuits.
Yet, for arguably the first time in history, “creative pursuits” are at a risk of being taken over by AI. Or are they? As with other sectors, the goal is to develop new ways to ensure that AI’s evolution in the creative industry complements human creativity and promotes ease of living. Some of these are already being applied today to solve different challenges and these include:
Music Therapy – AI can be used to create ethereal music that induces a state of restfulness. It does this by ordering music compositions and layering over acoustic features that can guide users to their desired state of ease. The compositions can self-learn based on the user’s preferences and be further finetuned to suit the listener’s tastes. This can be useful for terminally ill patients or
Content Generation – AI’s analytic capacity can be used to enhance the content creation process by conducting market analysis and social media sentiment analysis to generate content topics that human writers can further develop and publish on social media.
Music Auto-tagging – It is a constant challenge for streaming platforms to maintain well-curated music catalogues with the content explosion in today’s world. Inaccurately tagged music affects the recommender system’s ability to provide an immersive listening experience for users. AI algorithms can be used to automate the tagging process by categorizing music as required with accurate descriptive metadata (such as the artiste’s name and song title) and discovery metadata (such as the language, voice, genre, mood, energy, etc.) This feature saves time and helps musicians struggling to define their music identity.
Text-to-Image art – This application, described in the earlier section of this article, puts art creation within reach of everyone. A novelist, for example, doesn’t have to wait for an illustrator to design a book cover when they can generate one for themselves. Other algorithms are being developed to further finetune the text to image generation process by allowing for more specific images to be developed e.g. actual people and not just generic stock human images.