If artificial intelligence (AI) were a person, it would be one of the most talented and prolific creators of our time. AI’s creations can be found in a wide range of fields, spanning from literature and visual arts to fashion and gastronomy. While certain news outlets have chosen to focus on AI’s potential threat to human professionals and artists as a replacement, the truth is that AI is more of a tool than a creator in its own right.
As we start 2022, we must recognize the value AI brings to the creative arts. How we as a society create art and what we consider to be art have evolved. As we embrace new art in the forms of AR/VR and NFTs, we need to understand how we can use AI to transform the creative process and create art that better captures the comprehensive human experience.
AI as a content co-creator
AI applications in the creative fields started with language and text. We have all seen the headlines that read, “A robot wrote this entire article.” Using language models such as OpenAI’s GPT-3, journalists, authors and those experimenting with AI have been able to generate reams of writing with a few lines of input and a couple of clicks of the mouse. In March of 2017, writer Ross Goodwin embarked on a road trip with a camera, microphone and GPS unit hooked to his laptop to capture the inputs needed for AI to write the “next American road-trip novel.”
The idea was to take the sights and sounds on the road and feed them as the input to the AI’s neural network, and the output would be novel. The novel, which was published a year later, is being called the first novel was written by AI. While even Goodwin himself admits that the novel, 1 the Road, is not “a human novel, or anywhere near it,” it is a great example of how AI can co-create with artists in unprecedented ways.
In the field of music, AI has given us ”new” songs from deceased artists such as Jimmy Hendrix, Jim Morrison and Amy Winehouse. Using Google’s AI software Magenta, the Lost Tapes project studied 30 past songs from each artist to understand their songwriting styles and the typical structure of the songs to create new songs that bear the characteristics that are most representative of these artists. New artists can then perform these songs and bring them to life.
AI as a creative business partner
Regardless of the medium, AI can contribute to the creative process by processing more information than their human co-creators and finding patterns and trends that might otherwise be missed. It is the ability to analyze and derive insights that makes AI an excellent business partner, especially in the arts. The process of making art that is appealing to the masses has to do with understanding what is trending and even predicting it. AI can do just that, and this is where we expect it to continue to shine.
In 2017, Appier’s R&D team published an article in the Association for the Advancement of Artificial Intelligence publication about AI’s ability to generate fashion designs that performed well with human judges. Since the publication of this article, other real-life applications of AI in fashion have taken off as well. In 2019, two MIT students Pinar Yanardag and Emily Salvador developed a machine-learning software, Glitch, that designed a line of little black dresses. The AI software did not merely rely on fashion trends of the past for the designs, but rather, it used these large volumes of information to come up with new unique designs.
Nowadays, fashion brands are beginning to embrace AI’s predictive capabilities to forecast style trends that will influence the designs they make for upcoming seasons. The advantage here in using AI is that fashion houses can almost guarantee that their designs will be a hit because they have been made with what will be popular in mind.
In the culinary world, chefs can rely on AI to come up with food pairings and recipes never before imagined. This is the goal of projects such as IBM’s Chef Watson and Sony AI’s gastronomy project. Where the human chef may use their gut, the AI chef uses hard data and analysis for new delicious creations. What makes AI’s insights powerful is the years and volumes of data that serve as the basis for these recommendations. Add to it the experienced artist’s personal experiences and creativity, and we may just have great innovation in our hands.
Another strength of AI in creating art is that it can tailor the creation to a particular audience. For instance, imagine creating an online ad. There are so many choices to be made: what colour should the background be, what font should the words have, what images should be used? AI can help with these decisions because it understands what type and style of content will resonate with what audience. So, instead of one version of the ad, marketers can create multiple versions, use AI to learn through trial and error which one resonates the most with an individual and why, and in the future only show a particular type and style of an ad to the person.
A similar example can be found in video games. The fun of role-playing games is that there are so many different scenarios to explore. A decision can change the entire narrative of the game. With AI, game developers can make everything ten to a hundred times more complex. There are many more permutations that a decision can create, and each resulting scenario will feel as well-crafted and intentional as the main, intended storyline.
AI as an architect of the metaverse
The next frontier for Big Tech companies is the metaverse, for which AI will play a critical role in making it a reality. Artificial intelligence will be in charge of making these worlds within the metaverse feel as real as possible, and then making the process of creating new worlds more and more automated.
As mentioned previously, AI has strengths in personalization and creating multiple scenarios. These strengths will make it easier for developers of the metaverse to make worlds that are tailored to individual users. With AI, each world will continue to grow as the solution learns more about what each user likes and dislikes. Of course, the whole operation will take more than just AI; AR and VR technology, and the hardware and software components of each, will play a big role as well. Nonetheless, the creative and thus the human aspect of the metaverse will come from AI and the human developers that use it.
AI’s potential is truly limitless. The more input we feed it, the more and better output we get. Creative professionals are very aware of this. Instead of looking at AI as a competitor, many artists are choosing to see AI as a partner, something they can draw inspiration from and work with to better generate new forms of art.
Written by Dr. SD Lin, Chief Machine Learning Scientist and Dr. Min Sun, Chief Artificial Intelligence Scientist, Appier
