Johanna Reich, When all is quiet II.

From thought to form to pixel

The most exciting question about generative artificial intelligence seems to me to be the question of the development of consciousness: (when) will a generative AI develop its own consciousness?
A major irritation already arises today when using language models with chat functions: due to the text and language production of an AI, we humanize it and equate the process of algorithmically generated text with our human thinking.
To investigate this, I asked myself various questions: what does thinking actually mean? And how does this process begin? What precedes thinking? An image? A premonition? What happens in the moment before the thought can be clearly grasped?
The human mind works in a way that is strongly influenced by subjective experiences, emotions and intuitive understanding. Our thinking is not only logical and analytical, but also influenced by our memories, emotions and physical states. We process information in a broad context that is shaped by social, cultural and historical backgrounds. This leads to thinking that is flexible, creative and often imbued with prior knowledge and implicit assumptions.
In contrast, what we anthropomorphize as the “thinking” of an AI is based on clearly defined rules and algorithms. An AI processes vast amounts of data through machine learning and other techniques to recognize patterns and make predictions. While AI systems show amazing performance in analyzing data and solving specific problems, they lack the deeper understanding and contextuality that characterize human thinking. While AI can recognize patterns and draw logical conclusions, it currently lacks intuition, subjective experience and the broad context that shapes human decisions.
The difference between human thinking and the “thinking” of artificial intelligence is profound and concerns both the way in which information is processed and the fundamental conditions that enable this process. While human thinking takes place in a complex interplay of biological, emotional and physical factors, the “thinking” of an AI is based on purely algorithmic and data-driven processes.
In a self-experiment, I used an EEG to measure the concentration of my brain activity while I tried to think nothing until a thought manifested itself. I translated this data into a pixelated, slowed-down animation. At the same time, I captured the inner image that arises in my mind before thinking as an ink drawing on canvas. Both results can be seen as hybrid paintings in the series “When all is quiet”: the luminous animation of brain waves flashes behind the abstract traces of color.