> ## Content Index
> Fetch the complete content index at: https://aparaat.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# "Intrusive Thoughts" by Jurgis Peters
- URL: https://aparaat.ghost.io/reviews/
- Published: 2026-05-31T12:40:33.000Z
- Updated: 2026-07-29T07:18:16.000Z
- Author: Kristaps Brize
- Tags: reviews

Place: RIGA, RIXC Gallery Time: **May 28 – July 18, 2026** 

Artificial intelligence is frequently described as a *black box*—a system so complex that even its creators struggle to explain how particular outputs emerge. As increasingly consequential decisions are delegated to machine learning systems, this opacity has become one of the defining concerns surrounding AI. Rather than accepting that opacity as inevitable, Jurģis Peters' exhibition *Intrusive Thoughts* attempts to make the invisible visible.

![](https://storage.ghost.io/c/0e/99/0e998289-7490-4bd5-bd20-77bf6b71c2f0/content/images/2026/07/Screenshot-2026-07-28-073250.jpg)

Four displays and four gradual iterations of the initial image

The installation unfolds across multiple displays that reveal different stages of an AI system's internal processes. Beneath them lies the material infrastructure normally concealed behind sleek interfaces: exposed processors, dismantled hardware and tangled cables spread across the gallery floor. Peters deliberately refuses the illusion of immaterial intelligence. Behind every "thinking" machine lies an unmistakably physical network of silicon, electricity and computation.

A camera continuously observes the exhibition space, capturing visitors as they enter. Their image becomes the starting point for a recursive generative process. Passing through successive computational stages, the familiar scene gradually dissolves into increasingly saturated colours, heightened textures and abstract formations. Human figures slowly disappear beneath layers of algorithmic interpretation until almost no trace of the original remains. What begins as documentation evolves into an autonomous visual ecosystem that appears to feed upon itself.

Rather than simply demonstrating image generation, Peters exposes the visual biases accumulated within AI training data. Left without new human input, the system repeatedly gravitates towards recurring patterns—cellular structures, dot matrices and digital noise—revealing the aesthetic tendencies embedded in the model's statistical memory. The installation suggests that artificial intelligence is never neutral; it constantly reconstructs the world through the visual archetypes it has learned.

At a time when headlines oscillate between utopian promises and apocalyptic warnings about AI, *Intrusive Thoughts* offers a welcome alternative. Rather than asking whether machines will eventually surpass us, Peters asks a more immediate question: *What exactly are these systems seeing when they look at us?* By opening the black box—not completely, but enough to reveal its mechanisms—the exhibition replaces technological mystique with curiosity.