The first major social response to generative AI unfolded like a detective story. Schools sought software that could mark modeled essays, publishers tested graders for synthetic prose and watermarking platforms and “AI-generated” tags, as if the next age of trust would be won by sophisticated forensic tools.
However, AI detection systems are currently straining in two ways: they still produce both false positives and false negatives. Human-written work is regularly flagged as synthetic, while lightly edited AI-generated text often goes unnoticed. What’s worse is that tools that work well enough in one context may fail in another, either because of differences in language, style, model architecture, or triggering techniques. In practice, discovery now resembles an arms race, and arms races rarely end with one side declaring permanent victory.
This sounds alarming, but historically it is also known. When detection proves elusive, societies rarely resolve the crisis by endlessly refining forensic tools. Instead, they build systems of accountability through conventions of authorship, editorial oversight, provenance records, professional standards, and legal responsibility so that trust shifts from identifying provenance to establishing responsibility.
To put it more visually, it’s a shift from fingerprints to passports. Fingerprints imagine authenticity as something embedded within the object that can be revealed through closer inspection. Passports establish legitimacy through institutions, registries and verification systems.
Why discovery keeps failing
Discovery fails not because experts are stupid or expert systems are flawed, but because imitation improves faster than exposure and because authenticity is a social judgment as much as a technical one.
Art history has many examples to consider. Han van MeegerenForged Vermeers didn’t just fool casual viewers. They convinced serious authorities, only to have works approved by recognition later revealed as forgeries. The point is not simply that the experts were fooled; it was that seeing was never enough. Systems based on authenticity, derived from style, history and desire, become guidelines for counterfeiters to masterfully manipulate.
The pattern is repeated in Thomas ChattertonThe so-called “Rowley” poems, which were presented as medieval texts and welcomed as discoveries. Their reception depended on evidence as well as longing: a lost voice recovered, a past made present. The lesson here is that authenticity is not simply embedded in work, but is supported by institutions and incentives.
Generative AI scales that volatility. If a model can draft a believable essay in seconds and a human can polish it in minutes, the text becomes an unreliable witness to its origin: the “tell” changes and the cost of imitation collapses.
When faith moves: ancestry as a new center
Counterfeit currency provides a useful comparison. Modern economies do not rely on citizens to spot counterfeit banknotes. They rely on central banks, controlled issuance, security features, serial numbers, audit trails, regulated intermediaries and enforcement mechanisms. A banknote is trusted not because any person can detect a counterfeit, but because a system makes accountability possible. This is the logic of passports at scale, in which verification moves away from individual perception and into institutional infrastructure.
The art world learned the same lesson. After forgeries and contested attributions, markets shifted emphasis away from “best eyes” and toward provenance: ownership histories, archival documentation, conservation records, and chain of custody verification. Over time, paper trails often became more valuable than recognition.
HE is pushing faith in the same direction. Lasting trust will come less from stylistic residue than from provenance systems, including signed metadata, cryptographic authentication, and standards that allow platforms and publishers to confirm where content came from and what happened to it. Initiatives like C2PA and content credentials are early media passports. Their purpose is less “production detection” than “chain verification”.
Therefore, the question changes from “Can this be proven man-made?” for “Who released this, who handled it, and who is responsible for it?”
Authorship is a technology of accountability
Long before generative AI, societies learned to live with ambiguous origins, because authorship was never just a claim about who produced each word. It has always functioned as a practical accountability mechanism.
Consider ghostwriting. Political memoirs, executive books, and celebrity publications often involve uncredited writers and heavy editorial shaping. However, when a book contains an inaccuracy, the public rarely holds the ghostwriter responsible. Responsibility is attached to the name on the cover: the byline functions less as a forensic statement than as an anchor of responsibility.
Debates over Shakespeare’s authorship make the same point on a larger scale. We can never decide how collaborative the plays were or how much can be attributed to Shakespeare alone. But that is not the point. The works remain meaningful because societies can function despite imperfect knowledge of origin, as long as institutions stabilize responsibility and value.
Michel Foucault gave this mechanism a name: the author-function. Authorship, in his account, is a social role that classifies texts, regulates their circulation, and enables systems of ownership and punishment; he tells societies whom to praise, whom to pay, and whom to sue. It’s a point made even by Roland Barthes, who challenged the author as the sovereign source of meaning when he wrote that “The birth of the reader must be at the cost of the death of the Author.”
AI exerts pressure elsewhere, destabilizing the perpetrator as the node of accountability. If words can be generated, remixed and distributed at scale, the pressing question is governance: who is responsible for persuasion, defamation and synthetic expertise? In the age of synthetic generation, bylines, editorial practices, and publishing standards are not ceremonies, but rather infrastructures for trust.
(Re)production did not end the belief. Changed scaffolding.
The anxiety of authenticity is not uniquely digital, and the history of production (and reproduction) shows that belief usually survives by changing its scaffolding. Turning to examples of art, works from medieval and Renaissance workshops complicate our modern obsession with singular originality. In the studios associated with Raphael, Rubens or Titian, for example, multiple hands may contribute to a painting. Patrons bought the reputation of a workshop and the supervision of a master as much as a specific brush. Attribution continued because accountability was organized through commissions, contracts, standards, and reputational guarantees.
It’s a lesson we had to address with the rise of mechanical reproduction through photography. When the first daguerreotypes appeared, the images were treated as mechanical truths. But this truth was short-lived as manipulation quickly arrived through retouching, staging and darkroom editing. Photojournalism did not make every reader a forensic analyst; it built guardrails through editorial review, source discipline, codes of ethics, corrective practices, and reputational penalties. We learned to trust a process, not an image.
Walter Benjamin captured the change when he wrote that “what withers in the age of mechanical reproduction is the atmosphere of the work of art.” Reproduction changes what authenticity means and where value lives. Generative AI repositions that dynamic from reproduction to production itself. When reliable creation becomes cheap, the “aura” becomes harder to find within the artifact, so belief is re-aligned with what remains legible: who published, according to what standards, with what discoveries, and with what responsibility.
Borges and the collapse of origins as a discoverable property
Jorge Luis Borgesthe short story of Pierre Menard, Author of Quixote, has become an apt parable for authenticity and meaning in this age of AI. The story recounts how Menard reproduces parts of Cervantes verbatim, while the narrator nevertheless insists that the second text is “almost infinitely richer” because identical words have a different meaning when authored in a different era by a different person. Borges’s point is devastating to the fantasy of discovering authenticity, because the text itself cannot reliably reveal its origins; the origin is external, provided by attribution, context and institutional framing.
This is the AI problem in miniature. Two passages can be identical: one drafted by a student, one created by a model and edited by a human. If the object cannot hold a consistent internal signal of origin, looking more closely at it will not resolve the crisis of faith; the solution must be built around the artifact.
The institutional future after discovery
If detection remains unreliable, the most important change will not simply be that “fakes” flood the area, although they will. The most profound change will be institutional, a migration to systems of accountability that make trust enforceable even when its origins are unclear.
Therefore, we should expect more emphasis on provenance infrastructure, editorial accountability, and liability regimes that clarify who bears the risk when synthetic content causes harm. This is not utopian, as passports can be forged and institutions can fail. On the contrary, it must be recognized that systematic approaches scale confidence better than improving individual perception.
The future of trust in the age of generative artificial intelligence may depend less on identifying machine-generated content than on identifying who is behind it: because when authenticity becomes hard to see, responsibility becomes more valuable than discovery.
