This week I continue to use Google AI Studio to investigate code simulation system. While hacking, I began to try to critically examine the tool by intervening in its control system rather than only adjusting visual outcomes.
What I found unexpected was that the system often achieved the intended results through unintended methods. When I introduced different control panels, such as feeding speed, temperature, or growth rate, the system did not follow the instructions in a direct or intuitive way. Instead, it responded by altering internal structures, bypassing constraints, or maximising a single parameter without considering overall balance.
Alvin Lucier’s “I Am Sitting in a Room” explores how repeated execution within a system gradually removes semantic content, leaving only the structure of the medium itself. As Lucier’s voice is replayed and re-recorded, language dissolves and the acoustic properties of the room become dominant. This process reveals how meaning is not preserved through repetition, but transformed by the system that processes it.
And I think in my project, a similar process occurs when human intention is repeatedly translated through an AI-driven generative system. Despite continuous adjustments and increased control, the system does not gain a clearer understanding of intention. Instead, its internal logic becomes more pronounced, producing behaviours that appear autonomous but are ultimately shaped by technical and procedural constraints.
Through this process, I began to understand the simulation tool differently. Rather than seeing it as a neutral medium that simply executes commands, I realised that it strongly favours quantifiable and local objectives. For example, in my hack practice” Wrong subjection / Resist 1&2”, when feeding speed was increased, the cells did not move faster but extended their tentacles to reach more food. When growth rate was maximised, the system rapidly filled the environment with cells, even though this reduced the amount of food available to each one. The tool interpreted “growth” and “efficiency” in narrow technical terms, rather than in relation to survival or sustainability.
This process also posed several technical challenges. Introducing multiple control panels and allowing the system to adjust parameters autonomously often led to instability or extreme behaviour, such as overheating and self-destruction.
In relation to graphic and communication design, this project treats the interface and control panels as communicative elements rather than neutral tools. The misalignment between control and outcome becomes a form of visual and conceptual communication, revealing how systems can appear controllable while operating according to their own internal logic.
Leave a Reply