The finalists have been announced for the AI Artathon, a first-of-its kind international event designed to highlight the creative potential for humans and machines to work together to create art.
More than 2,000 people from more than 50 countries — including artists, graphic designers, experts in artificial intelligence, and programmers — applied for the competition, which was organized by the Saudi Data and AI Authority (SDAIA) as part of the Global AI Summit.
After a qualifying round in Riyadh in January during which the participants competed to create initial AI-based artworks, 300 of them were selected to progress to the next stage. They formed 20 teams that took part in a boot camp, during which they were taught by experts to use in their work machine-learning algorithms and generative adversarial networks (a framework for machine learning in which two neural networks compete against each other).
The 10 teams that have reached the final were announced on Tuesday during a live-streamed online event that showcased the artworks created by the teams using AI tools and techniques. The top three, who will share a $133,000 prize fund, will be announced on Oct. 22, the second and final day of the Global AI Summit.
The online event will bring together key decision-makers from government, academia, business and technology to discuss global issues and challenges, and the ways in which AI can benefit humanity and shape the new normal.
“SDAIA is proud of these high-quality artworks created by the finalists of the AI Artathon,” said Abdullah Al-Ghamdi, president of SDAIA and chairman of the organizing committee of the Global AI Summit. “I am very pleased to see artists and AI developers collaborating to utilize the capabilities of advanced AI to create these innovative artworks, despite the novel coronavirus pandemic. I look forward to announcing the final winners during the Global AI Summit.”
The final 10 teams competing for the top prizes are:
The idea for their artwork came from the question ‘what if Picasso had been Saudi?’ The team created a database of works by Picasso and Saudi artists that the AI technology used to create an artwork highlighting the depth, richness and diversity of Saudi art.
In an effort to reflect the beauty of the nature in the Kingdom, the artists used geographic and topographic images, along with existing lines and shapes, to train the AI to combine them and create a digital artwork based on new patterns and inspirations.
Maruem — The Chimera Ant Queen
Inspired by popular manga and anime series “Hunter x Hunter,” and in particular characters called Chimera Ants, this project combines digital graffiti with other artworks and concepts to create a unique new piece that traces the development of a Chimera Ant from single cell to birth.
AI Authentic Clay
The development of personality is usually the result of accumulation of knowledge, including, for example, awareness of culture, religion and civilization. The members of this team decided to express their deep feelings about the authenticity of their culture, presenting an experience that combines and reorganizes shapes and patterns through the use of AI.
Esoteric & Energetic “Heya”
Words can have a number of meanings, energies and inspirations. For example, “heya,” which translates as “she” in English, can have many meanings: “she” might be a daughter, a sister, a mother, a hard worker, or many other things. This artwork is inspired by the energy, feelings and strength conveyed by “heya.” The team collected selfies of diverse women and drawings reflecting the word “she,” and AI technology was used to analyze and link them to create a new digital work of art.
The Beauty of Belonging
Architecture is a fundamental characteristic of any cultural production that enhances the sense of belonging, so this artwork aims to shed light on the richness and variety of architecture in Saudi culture. Different architectural styles and patterns are combined and integrated using AI algorithms to produce an artwork with a new color and style, creating an abstract visual experience.
Museum of GANs
What if we could depict the emotional states and impressions of a character based on portraits? And what if we could create new artworks based on differing emotional states? This work tries to answer these questions by collecting and classifying a large number of well-known portraits according to their emotional states and impressions, which the machine used to create the artwork.
As humans, when we read a text there are always intertextualities and associations to perceptions, figures and so on in our minds. To create this artwork, the team developed a model that enables the machine to read these associations using AI. A random text generator was used to create stories that are used to transform the textual into the visual.
Different architectural shapes and styles have developed over time in the Arab world that are used in buildings, mosques and other landmarks to create characteristics unique to Arab art. The team combined many different patterns to develop a single aesthetic pattern that offers the best comprehensive picture. Patterns were processed using deep convolutional generative adversarial networks, which classified, connected, developed and redrew shapes the machine used to create a new work of art.
What distinguishes digital art created using AI is its limitless capabilities to recognize and translate human creativity. Taking the novel as a crucial element of literary creativity, this work is the result of an attempt to inspire the computer to create art by reading a novel using linguistic programming, extracting its concepts, and then transforming those concepts into photos and shapes.