Quality Assurance for AI-Mediated Work
Quality assurance for AI-mediated work names the institutional design problem of building evaluation systems that penetrate the smooth surface of machine-generated output to assess the comprehension beneath. The historical precedent is the editorial function in publishing, which evaluated manuscripts for substantive quality rather than surface polish. An editor at a reputable press in the seventeenth century did not merely check that the text was legible; she evaluated the argument, assessed evidence, identified weaknesses, and determined whether the manuscript met a standard of quality that the press's reputation required. The AI equivalent must operate at scales orders of magnitude larger than its historical antecedent.
In the AI Story
The most dangerous feature of AI-generated work is the independence of its surface quality from its substantive quality. The code compiles regardless of whether its author understands the architecture. The legal brief cites relevant precedent regardless of whether its author has read the cases. The medical recommendation follows clinical logic regardless of whether the requester can evaluate the reasoning. A quality assurance system designed for AI-mediated work must penetrate the surface to evaluate the substance, and it must do so at a speed compatible with the production rate of the tools it evaluates.
The scale challenge is acute. The production rate of AI-assisted work exceeds the production rate of the printing press by a corresponding factor. Whether AI systems can themselves be designed to evaluate AI-generated output for substantive quality rather than surface coherence is an engineering question the current state of the technology leaves open. Whether human evaluation can scale to match AI production is an institutional question the current state of professional practice also leaves open.
Several partial solutions are emerging. Peer review structures are being adapted to AI-assisted work. Professional certifications are being revised to assess comprehension alongside output. Liability frameworks are being extended to hold practitioners accountable for AI-generated work regardless of authorship. Each of these is a dam; none of them is yet at the scale the problem requires.
The Cipolla framework's contribution is diagnostic clarity about why the question is urgent. Without substantive quality evaluation, the smooth output of AI-assisted production will conceal the comprehension gap that is the primary mechanism by which the technology amplifies damage. The absence of the evaluative infrastructure is not a feature that resolves itself; it is a structural deficit that compounds until the institutional response arrives or the damage becomes unmistakable.
Origin
The concept extends Cipolla's analysis of editorial institutions in early modern publishing to the AI context. Contemporary work on AI governance by Fukuyama, Allen, and others has developed specific regulatory proposals; the problem's scope remains under-theorized.
Key Ideas
The surface-substance gap. AI-generated output's polish is structurally independent of the comprehension that directed it.
The editorial precedent. Publishing's quality assurance function offers a historical template, scaled for a production rate orders of magnitude higher.
Partial solutions emerging. Peer review, professional certification, liability extension — each a dam, none yet at required scale.
Diagnostic urgency. The structural deficit compounds until resolved; it does not self-correct.
Appears in the Orange Pill Cycle
Further reading
- Carlo Cipolla, Before the Industrial Revolution (Norton, 1976)
- Elizabeth Eisenstein, The Printing Press as an Agent of Change (Cambridge, 1979)