Why draw at all
Every case in the first series is already documented in prose, often at great length. The Royal Commission into Robodebt runs to close to a thousand pages. The Dutch parliamentary inquiry into the childcare-benefits affair is a book. A reasonable question is what a drawing can add to a record of that size.
The answer the work proposes is about relation. A report describes harm one dimension and one period at a time: first the debts, then the health effects, then the court cases. The spectrogram puts all seven dimensions of harm on one time axis, so that the viewer sees at once what the report can only list. When the system started to harm people, which kinds of harm came first, when someone intervened, and what remained afterwards. The gap between the first warning and the effective intervention, which in Robodebt is two and a half years, is visible as a distance before it is read as a number.
The Encounter Plate answers a different need. Each automated system in the series observed people: it scored them, ranked them, flagged them. Harun Farocki called images made by machines for machines, which are not meant to be looked at by people, “operational images” (Farocki 2004); Trevor Paglen extended the idea to the infrastructures that now produce most images in the world (Paglen 2014). The plate reverses the direction of that look. It draws the system as an object under observation, in the visual language of an astronomical photograph, and each plate carries the prefix OBS. in its title. The eye that scored people becomes the thing being recorded.
A lineage
The practice belongs to a line of art that works with public records. Hans Haacke’s Shapolsky et al. Manhattan Real Estate Holdings, a Real-Time Social System, as of May 1, 1971 consisted of photographs and documents drawn entirely from public property records. The Guggenheim cancelled the exhibition before it opened. The point of Haacke’s method was that the material was public and verifiable, and that its arrangement made visible a pattern that the scattered records concealed.
Forensic Architecture, the research agency founded by Eyal Weizman at Goldsmiths, has turned that principle into a methodology that produces evidence used in courts, parliaments and truth commissions (Weizman 2017). Susan Schuppli’s work on the “material witness” describes how objects and media can carry traces of the events they were part of and can be read as testimony (Schuppli 2020). The present work is smaller in scale and makes a narrower claim. It does not produce new evidence. It takes evidence that institutions have already established and keeps it in a form that can be compared across cases and over time.
Four decisions
Every mark comes from the record. Each visual property of a plate or spectrogram is set by a coded value in the case record, according to a written specification. The same record always produces the same image, and an image changes only when the record changes through a documented revision. Sol LeWitt wrote that in conceptual art “the idea becomes a machine that makes the art” (LeWitt 1967). Here the rule serves a different end. It is not there to free the artist from decisions about form; it is there so that every form can be checked. Anyone with the record and the specification can verify why a band is bright or a ring is blurred.
Harm is never reduced to one number. Each case is coded on seven dimensions: liberty, dignity, employment, family, housing, health and reputation. The scores are never summed or averaged, and there is no ranking of cases by total severity. The decision is deliberate. Every system in the series did the opposite: it compressed a person’s circumstances into a single risk score and acted on that number. A record that answered with its own single number would repeat the operation it documents. D’Ignazio and Klein (2020) describe the choice of what to count, and how, as a question of power before it is a question of technique. In this method the refusal to aggregate is the main way that choice is expressed.
Uncertainty is drawn. Where the evidence is too thin to score a dimension, the band is hatched. Where harm is suspected but cannot be measured, the band has a dotted outline. Where the causal chain is uncertain, the ring on the plate is blurred. Where the outcomes are estimated from a mechanism rather than counted, the mark of irreversible harm is hatched instead of solid. Weizman describes much contemporary violence as taking place at “the threshold of detectability”, where it is hard to prove either way. A visual record that showed only what is proven would draw these cases as less harmful than they are. One that filled the gaps with estimates would claim more than it knows. The method marks the gap instead, and says what kind of gap it is.
The people affected are not pictured. No plate or spectrogram contains a face, a name or an identifying detail. People appear only as stars, and the brightness of a star does not stand for the harm done to any one person, which is not known. On the Optum plate the stars are fainter and more numerous, because the patients affected never learned that a score had been applied to them. People are named in the case texts only when they have made their own case public. Children are never identified. Susan Sontag asked what it means to look at images of other people’s suffering, and what such looking does for the viewer (Sontag 2003). This work does not make such images. It asks the viewer to look at the system instead.
Why no generated images
The work uses no AI-generated imagery. The rule is not a position on generative tools in general. It follows from the purpose of the record. An image made by a generative model cannot be traced back to its causes: the same prompt produces different outputs, and the model’s training data cannot be audited from the image. A work about decisions that could not be traced back to their reasons would be weakened by using a process with the same flaw.
Limits
Coding is judgment. Whether the dignity harm of a scoring system deserves an 8 or a 7 is a decision made by a person reading the evidence, according to written criteria. The method does not hide this. The codebook, the scoring criteria and the sources for each case will be published in an open catalogue, separate from the artworks, so that the coding can be examined and disputed. Each case record states when it was coded and when it was last verified, and corrections are dated. A visual record that asks institutions to be accountable for their automated decisions has to meet the same standard for its own.
References
- D’Ignazio, C., and Klein, L. F. (2020). Data Feminism. Cambridge, MA: MIT Press.
- Farocki, H. (2004). Phantom images. Public, 29, 12–22.
- Haacke, H. (1971). Shapolsky et al. Manhattan Real Estate Holdings, a Real-Time Social System, as of May 1, 1971. Photographs, documents and maps.
- LeWitt, S. (1967). Paragraphs on conceptual art. Artforum, 5(10), 79–83.
- Paglen, T. (2014). Operational images. e-flux journal, 59.
- Schuppli, S. (2020). Material Witness: Media, Forensics, Evidence. Cambridge, MA: MIT Press.
- Sontag, S. (2003). Regarding the Pain of Others. New York: Farrar, Straus and Giroux.
- Weizman, E. (2017). Forensic Architecture: Violence at the Threshold of Detectability. New York: Zone Books.