Open post

AI Dreams

This post is about visiting MCA Data Dreams:

Art and AI

Location MCA Sydney Harbour at the Rocks

Note to the reader

I have published here asking Claude AI about MCA and my own photos. So the replies of Claude AI were carefully read and re-edited to publish this blog.

Credit MCA | phb

Anatomy of an AI System (2018) — Crawford & Joler at Data Dreams, MCA Sydney

What the installation actually is

The work consists of a large-scale map — a visual essay requiring minimum print dimensions of 220×360 cm — that uncovers the invisible matrix of human labour, energy consumption, and resource extraction hidden behind digital networks and AI. It maps the full chain of manufacture behind the Amazon Echo, from geological extraction through data exploitation and energy consumption during AI training, to the device’s death and disposal.

At the MCA, the installation went beyond the map. Situated next to the diagram is a cabinet containing samples of the rare minerals required to produce the Echo, each accompanied by a label noting its industrial application, the site of its extraction, its level of toxicity, and the illnesses associated with exposure. And of course the dissected Echo device itself lies there — opened, gutted, made literally transparent.

The core argument — your cui bono lens applies directly: Who pays here what exactly?

Who profits most?

Crawford and Joler’s research critically examines the concept of the “cloud” — a metaphor that evokes a disembodied, ethereal image. In reality, digital technologies are inextricably tied to material resources and economic and geopolitical interests. Countless human workers are needed to construct, train and maintain the system: to mine minerals, produce hardware, generate training data and manually correct the AI — often in poor, underpaid, health-damaging conditions. And whilst the consumer is analysed in all areas of life, practically transparent to corporations, the companies themselves obscure information about themselves. The true costs — social, ecological, economic, political — remain hidden.

Crawford and Joler make the Frankfurt School’s Ideologiekritik of Horkheimer & Adorno fully visible here. The “smart home device” is the commodity fetish par excellence: it conceals its own conditions of production behind a friendly cylinder and a voice called Alexa.

As Crawford and Joler observe, the user performs simultaneously the function of consumer, resource, worker, and product. As a consumer they receive convenience; as a resource they provide their voice to a large dataset; as labour they provide feedback; and as product they contribute to development. Marx would have recognised this immediately — surplus value extracted at every node, including the most intimate one: your own home.

The deep time dimension — the most unsettling move

This is where the installation achieves something genuinely philosophical. The minerals in that cabinet have rested in the earth for vast geological aeons, only to be torn from their strata for the manufacture of a device whose life cycle is shockingly — even offensively — brief by comparison.

Each object in the extended network of an AI system, from network routers to batteries to microphones, is built using elements that required billions of years to be produced. Looking from the perspective of deep time, we are extracting Earth’s history to serve a split second of technological time, in order to build devices often designed to be used for no more than a few years.

The broader exhibition context

What marks Data Dreams out as important is the extent to which so many of the works push back against disembodied narratives. The exhibition simultaneously rematerialises and demystifies, returning discussions of AI to the worlds on which it depends. It offers a slowing of pace — an invitation to look carefully at what AI takes, what it costs, and what kinds of lived realities it is already shaping.

The exhibition opened 21 November 2025 and runs until 26 April 2026.

Location Sydney Harbour under the iconic Harbour Bridge.

Credit MCA ~ phb

There is something quietly ironic about the fact that you photographed this label on your phone — a device whose own mineral cabinet would look remarkably similar. Crawford and Joler would appreciate that. The viewer is always already inside the diagram.

The cui bono is answered explicitly and globally: the extractive infrastructure of AI serves the platform monopolists, at cost to miners, data workers, ecologies, and — in deep time — the planet itself:

Mother Nature and the planet:

Credit MCA | phb


Image 1 — The wall text: “Planetary Costs: Environmental Impacts of AI / What Does AI Cost the Planet?”
The curators’ question is deceptively simple, but it encodes the entire Frankfurt School critique of instrumental reason. Notice the rhetorical structure: “Who bears the environmental risks, and who benefits from the convenience?” — that is your cui bono in exhibition-catalogue prose. The asymmetry is the argument. The ones bearing the risk (miners in Congo, data labellers in the Philippines, communities near server farms) are never the ones enjoying the convenience.
The second paragraph is philosophically the richer one: “Thinking about AI ecologically means tracing connections — between server farms and rivers, batteries and mines, code, and climate.” This is systems thinking in the Humboldtian tradition — everything connects, nothing is isolated — applied to the most aggressively isolationist technology marketing discourse of our era. Big Tech sells transcendence; Crawford and Joler sell geology.
The final question — “What counts as a meaningful image or output when each has an environmental footprint?” — is quietly devastating. Every generated image, every AI chatbot response, every query to a large language model carries a carbon and water cost that is structurally invisible to the user. Including, one must add, this conversation.
Image 2 — The installation itself: the great map on lightbox, with mineral samples at the base
The photograph captures something the published image doesn’t quite convey: the scale and darkness of it. White lines on black — it reads like a technical blueprint, a circuit diagram, a medical anatomy chart, all simultaneously. The three vertical columns map the three extractive axes: material resources (left), computational processing (centre), and data/disposal (right).
What your photo shows beautifully is the transition at the bottom: where the abstract white-line diagram literally meets the physical mineral samples sitting on a shelf in front of it. That is the conceptual hinge of the entire work. The map is intellectual; the minerals are real. Ochre-red ore, black rock — billions of years old, sitting below a diagram of a device with a two-year lifecycle. The juxtaposition is not commentary. It is evidence.
The lightbox format is a precise choice: it glows, it illuminates from within, it has the aesthetic of a radiology scan. You are looking at the body of a system that was never meant to be seen.

The Installation is an example of the Power of Art Work:

The artist (A) and his work (C) and the visitor (B):

The Golden Triangle again.

Credit MCA | phb
Minerals & AI
Credit MCA | phb
Credit MCA | phb
Scroll to top