Molar Security
Type: Hybrid Media Video Essay
Specs: 16’00” color, sound, single-channel
Year: 2025
01. 概述 / Synopsis
Molar Security is a Hybrid Media Video Essay that fuses AI-generated user profiles with military targeting interfaces to critique the politics of algorithmic warfare and data commodification. It reflects on digital complicity, emotional extraction, and the blurred boundary between security and surveillance.
I provided my social media feeds and browsing records to AI system, allowing it to generate a log based on my online behavior. This AI-generated persona, composed of my residual digital traces, was then transformed into speech through AI voice-cloning, merging virtual identity with noise-generated-voices.
In an AI-driven landscape, this work examines Israel’s military use of artificial intelligence, which generates target lists by analyzing phone signals, social media interactions, and satellite imagery. By combining screenshots and interfaces from actual military AI targeting systems, such as “The Gospel,” I explore these inherent biases, double standards, and contradictions within AI warfare systems, which are designed to “tolerate a degree of collateral damage”.
02. 堆栈 / Automated genocide
The instrumentalizing of information turns the whole planet as a rationalized sphere of resource extraction. Dominators power may take the form of a infrastructure that enables them to extract information asymmetries from both capital and from subordinate classes and to accumulate asymmetric information about all of these activities.
LLM (a guessing machine), makes consequential decisions that affect the lives of people under military occupation. In this video, I used the names of three interconnected military AI systems “Alchemist,” “The Gospel,” and “Fire Factory” as structural components, integrating my personal data and sound synthesis to speculate in reverse about the opacity and biases of this “black box” system.
I refer to Benjamin Bratton’s “Stack Theory” to analyze AI weapon systems across planetary-scale computational layers, from the Earth’s energy-consuming infrastructures to algorithmic cloud processing, urban destruction, and user-level decisions:
Figure: Analytics Diagram
- Earth Layer: Physical infrastructure supporting operations (data centers, satellites).
- Cloud Layer: AI processing of data for target lists.
- City / Address Layer: Utilizing GPS coordinates to locate target buildings.
- Interface / User Layer: Military operators making decisions based on AI recommendations.
I compare this analysis with my personal experiences on digital platforms, where emotional vulnerabilities are converted into data and sold back as illusions of security - comfortable but fragile bubbles carefully maintained by algorithms.
03. 音频与合成 / Audio Elements
For this video, I used an experimental sound-synthesis app “sketch synth” inspired by Iannis Xenakis’s UPIC system. This app can transform hand-drawn sketches into synthesized sound, converting visual imagery into audio. By integrating this method, I explored the interplay, adding another layer to simulate the drone’s view.

Through this project, I confront my own complicity as a provider of “surplus data” within the system, raising questions about agency, safety, and ethics in the AI-mediated world we inhabit.