__VA_OBS__

Currently, every several months, with ever-increasing frequency, a new AI system is launched, attracting worldwide attention. The growing use of systems such as ChatGPT, Adobe Firefly, MidJourney raises an issue regarding the control of the individual’s products of the interaction with them, and their personal relevance. Despite their extensive knowledge in a variety of areas, these systems have no information about the specific individual using them. Upon receiving an order, they run an algorithm as if trying to solve a riddle and offer a final and non-negotiable answer. This process also makes us relate to the process from a binary perspective. 

In order to find a solution for this problem, I asked myself during my research, “How would an intimate acquaintance between a person and an AI system affect creative processes?” I tried to answer the question from a responsible point of view, that leaves the agency and most of the control over the outcome in human hands, positioning these “omniscient” systems as a consultation and supportive tool. 

__VA_OBS__ redefines the interaction between the artist and an AI system, in order to channel the system’s extensive information into deep acquaintance with the specific person with whom it interacts. 

The system will interview artists to attain in-depth insights into their motivations, thereby deconstructing the creative process. With ongoing usage, the system will hone itself and channel the conversation to new and surprising areas that will help the artists hone their own approach. Based on preliminary learning of the artists, the system studies their unique artistic language, and offers visual references that can be viewed in a less adversarial manner – not only terrible failure or resounding success – enabling the artists to discover inspiring elements within them. That interaction harnesses AI’s ability to quickly produce a variety of outputs to human creativity, without which it could not have existed. 

In my project, I have chosen to focus on the creative worlds and use them to present what individually tailored AI systems can enable. This raises questions and future possibilities in a variety of other areas where similar interventions may be undertaken. I hope that future AI products will implement the approach proposed herein.