The warning for architecture is not simply that artificial intelligence can produce poor writing. It is that automated text can make claims about buildings without anyone having seen them, weakening the link between criticism and the places it describes.

A test involving Victor Hugo’s phrase “Ceci tuera cela” illustrates the problem. Asked to explain “This will destroy that”, chatbots offered the familiar interpretation: printed books would displace buildings as records of human thought. But Hugo gave the line a second meaning, too. In Notre-Dame de Paris, he presents the printing press as a force that could undermine the Church’s authority. The excerpt from the novel published by Reading Design makes clear that the phrase concerns both a change in how knowledge is recorded and a challenge to who controls it.

That history offers a more complicated lesson than one medium simply replacing another. Mario Carpo’s Architecture in the Age of Printing, published by MIT Press, traces how communications technologies shaped architectural theory. Printing altered how architects could share images, models and ideas; it did not make architecture itself disappear. Each new medium changes the ways knowledge travels, and what people can do with it.

The same question now applies to digital systems. Architectural knowledge survives in buildings and models, in people and institutions, in physical records such as drawings and photographs, and in digital files. Language models draw on digital material, but produce new text by identifying patterns in their training data. The US National Institute of Standards and Technology warns that generative AI can present false or inconsistent material with confidence. A fluent answer, then, is not necessarily a complete or dependable one.

The risk for criticism is that a chain of automated summaries and generated copy can circulate without anyone returning to the building or checking the account against first-hand experience. Luciano Floridi’s idea of “distant writing”, as summarised in the supplied paper, describes a human author shaping a narrative through prompts and revision while a language model generates the prose. That may leave room for human direction, but it also raises a question about what is lost when the writer’s work moves further from direct observation.

The answer matters because visiting a building is not just a way to gather details. It is part of how a critic develops judgement: noticing what photographs omit, testing claims against the site and deciding what deserves attention. The essay argues that automating routine writing may also remove the apprenticeship through which writers learn those skills. The concern is not that every use of AI produces worthless work, but that a system can produce text without having witnessed what it describes, and without being answerable for its conclusions.

The word “slop” has come to name one version of this anxiety. Merriam-Webster’s announcement of its 2025 Word of the Year defined it as low-quality digital material usually generated in quantity by AI. The term points to volume as well as quality: a flood of plausible content can make it harder to distinguish considered judgement from copy assembled for speed or profit.

Yet the history of print cautions against assuming that a new medium only destroys what came before. As Hugo’s two-part interpretation suggests, the press changed both the circulation of ideas and the institutions that claimed authority over them. Carpo’s account likewise shows how printed images helped reshape architectural thinking. The question is not whether architecture can survive another communications shift, but how its knowledge will be tested and transmitted through it.

For the essay’s author, the answer rests partly with criticism that can be attributed to people and institutions. A magazine can commission writers, take a position and be held responsible for it; a dated print issue also preserves a fixed version of an argument. A chatbot, by contrast, can generate a fresh response each time, without a lasting reputation or a record of judgement to defend. The distinction is not a claim that print is infallible, but that accountable criticism depends on identifiable witnesses.

Architecture will continue to be built on sites, while its ideas and histories circulate through changing media. Generative systems may assist that circulation, but they cannot replace the act of going to see a building and judging it in context. The essay’s central warning is that criticism risks losing its grounding if the path from evidence to opinion is reduced to a loop of digital material and automated prose.

Source Reference Map

Inspired by headline at: [1]

Sources by paragraph:

Source: Noah Wire Services