ARTificial: Exploring the Boundaries of Artificial Intelligence and Art

An Autonomous Robot Redefining the Concept of Art Creation

This robot, aptly named ARTificial, is among the first fully autonomous machines that create art with barely any human authorship. Designed by Titus Ebbecke, this groundbreaking installation challenges the belief of art creation as a property of human or artificial intelligence. Upon starting the system, ARTificial generates a completely new portrait with the help of artificial intelligence. A novel algorithm converts the digital AI art into coordinates, and the robot then paints it on canvas with acrylic. The sculpture works without human supervision, pushing the boundaries of what we consider art and the role of the artist.

The uniqueness of ARTificial lies in its ability to create art without any human intervention. The images are generated using a generative adversarial network, a cutting-edge technology that combines machine learning and robotics. The algorithm, written in Python with OpenCV and NumPy, allows the robot to paint unlimited images. The KUKA robot, weighing 450kg and measuring 3000 X 2000 X 1500mm, is connected to a computer via an Ethernet connection. The custom-made foundation, constructed with wood, concrete, and steel, ensures the safe operation of this industrial robot.

The key element of this project is its autonomy. ARTificial decides what to generate on a random basis and paints as it wishes, attempting to emulate human artists and their creative freedom as much as possible. With the push of two buttons, the robot starts generating a random image, takes paint, cleans its brush, and continues to paint until the artwork is finished. This process, devoid of human intervention, raises questions about the nature of art and the role of the artist.

The creation of ARTificial was not without its challenges. Contacting a robot manufacturer, securing funds, transporting the robot, and installing its numerous control interfaces were just some of the hurdles that had to be overcome. Additionally, a foundation had to be designed and constructed to ensure the safe operation of this dangerous piece of industrial equipment. The author also had to learn how to program a robot, train the artificial intelligence for multiple days on a powerful GPU, and develop a completely novel algorithm for converting an AI image into coordinates in a safe manner.

This bachelor thesis project, realized from February 2019 to July 2019, demonstrated its ability during a two-day exhibition in August 2019. Over the course of the exhibition, ARTificial painted for over 10 hours in front of a hundred visitors, captivating them with its unique artistic process. The research conducted for this project, in cooperation with KUKA and supervised under the head of corporate research, delves into the intersection of artificial intelligence, machine learning, and robotics. The algorithmic basis of the image-to-coordinate translation will be published in 2020, contributing to the growing body of knowledge in this field.

ARTificial's groundbreaking design and its exploration of the boundaries between artificial intelligence and art have not gone unnoticed. In 2020, it was awarded the Bronze A' Design Award in the Robotics, Automaton, and Automation Design category. This prestigious award recognizes designs that exhibit strong technical and creative skills, incorporating best practices in art, science, design, and technology. ARTificial's ability to generate non-human artworks, select random motifs, and paint them on physical canvas without any human intervention showcases the potential of machines to become artists, challenging our preconceived notions of creativity and artistic expression.