Trust Is a Technology
You take a prescribed medicine without checking it, you sign a deed without redoing it. That is not naivety: it is the most economical technology ever invented. AI has just short-circuited it.
Every day, you perform acts of vertiginous trust. You swallow a pill because a person you barely know prescribed it. You sign a deed of sale without having verified a single rule it applies. You board a plane flown by strangers, designed by engineers you will never meet. At no point do you verify, and it is neither laziness nor credulity: it is the normal functioning of a system that verified for you. The diploma verified competence, the professional body verified admission, the oath bound publicly, discipline watches continuously. You do not trust a person; you trust a social machine that made verification unnecessary.
For that is what it is: a machine, a technology in the proper sense, a device built to solve a problem. The problem is the oldest of human cooperation: how to deal with strangers without spending one's life verifying? Verification is the hidden cost of every transaction, and a society where everyone had to verify everything would be a paralyzed society. Professional trust is the solution societies invented: concentrate verification at one point, once, at the entrance, to make it unnecessary everywhere else, at every act. A physician is verified once, at admission; society is spared verifying each of their prescriptions for forty years.
Trust is not a sentiment. It is the most economical technology societies ever invented.
The cost no one calculates
To grasp this technology's value, one must do a calculation no one does, precisely because trust spares us from it. What would life cost without it? Every medical consultation would require a counter-expertise; every legal deed, a second independent review; every certified balance sheet, a new audit by its reader. The cost of generalized verification would quickly exceed the cost of the work itself. Economists have a name for this friction: transaction costs. Societies that reduced them prospered; those that did not saw every exchange bog down in distrust. Professional trust is, by this measure, infrastructure in the same sense as roads: invisible when it works, and foundational to everything that moves on it.
This infrastructure has a remarkable economic property: it is expensive to build and almost free to use. Building trust in a profession took centuries: establishing the training, founding the bodies, codifying the ethics, instituting the sanctions, and above all accumulating the proof, generation after generation, that the system keeps its promises. But once built, it is used freely, at a glance: the title suffices. "Avocat at the Paris bar," "physician on the register" are compressed signals that carry, in three words, centuries of institution. That is the mark of great infrastructures: an immense investment, amortized over unlimited use.
Trust costs centuries to build and nothing to use. That is the mark of great infrastructures.
What the title really says
One must be precise about what the signal carries, for that is where everything will play out. The professional title does not say "this person never errs." It says three subtler things. First: this person was selected and trained by criteria you need not know, by people competent to judge them. Second: they operate under a regime of rules and surveillance that makes fault costly to them. Third: if they fail, there is recourse, a forum, a remedy. The signal does not promise perfection; it promises a system around imperfection. And that is why it authorizes not verifying: not because error is impossible, but because it is contained.
Notice that nothing in this signal bears on the content of the work. The reader of a legal opinion cannot, by definition, assess its correctness: if they could, they would not have needed the jurist. Professional trust solves precisely this paradox, the most twisted of all: how to rely on work one is incapable of judging? Answer: by relying not on the work, but on the apparatus around its author. It is a second-order trust, carried by institutions and not by examination. The entire economy of knowledge rests on it, for all specialized knowledge is, by nature, unverifiable by those who need it.
Nor did this system fall from the sky, and its history speaks its value. Organized professions were born, almost everywhere, from a crisis of trust: charlatans in medicine, forgers in accounting, adventurers in law. Each time, the same movement: serious practitioners agreed to bind their own hands, to submit to a selection, an ethics and a discipline they imposed on themselves, in exchange for which society granted them what it grants no one else: the presumption of reliability. The oath is not folklore; it is the trace of that contract. The professions bought the public's trust at the price of their own liberty, and it is that price paid that makes the signal credible.
The title does not promise perfection. It promises a system around imperfection, and that is what spares verification.
The short circuit
Now here is what artificial intelligence introduces into this machinery. It produces, in unlimited quantity and at zero cost, content bearing every appearance of professional work: the structure, the tone, the technicality, the assurance. This content imitates the signal without having passed through the machine. No selection, no oath, no discipline, no recourse stand behind it; but nothing on its surface distinguishes it from work that comes with them. Yet the whole technology of trust rested on a tacit hypothesis: that professional form was hard to produce, and therefore signaled, fairly reliably, passage through the apparatus. That hypothesis has just fallen. The plausible has become free.
The consequences unfold in a chain. If form no longer signals the apparatus, the reader can no longer rely on it, and the cost of verification, which trust had abolished, returns to their charge. But it returns at the worst moment: precisely when the volume of content explodes, making systematic verification materially impossible. Worse: the doubt does not stop at generated content. Since one can no longer distinguish, all content of professional appearance becomes suspect, including that which really did pass through the apparatus. Authentic work pays the price of free imitation. It is the classic mechanism of currencies: counterfeit never circulates alone, it depreciates the genuine.
Everyone can already feel this shift in their own day. Before an impeccably drafted brief, a question now intrudes that did not arise three years ago: who really wrote this, and what stands behind it? The text has not changed in quality; it is the link between quality and guarantee that has broken. This new question, asked before every document, is the exact symptom of the short circuit: verification, which the professional signal had made unnecessary, becomes a reflex again. And a generalized reflex of verification is precisely what the technology of trust was invented to abolish.
The plausible imitates the signal of trust without passing through the machine that makes it. And counterfeit always depreciates the genuine.
The inflation of verification
The first effects of this inflation are already visible. Organizations reinstate double reviews they had abandoned; courts require certifications of origin for filings; newsrooms verify content they would have published without a thought five years ago. Each of these reactions is rational, and their sum is an impoverishment: society is relearning to pay a cost its institutions took centuries to abolish. If nothing opposes it, the trajectory is that of an economy of distrust, where every recipient redoes part of the sender's work, where exchange slows, and where the advantage goes not to the most competent but to the most verifiable.
But that last phrase contains, on reflection, the whole reversal. If the advantage goes to the most verifiable, then scarcity has changed sides. What was scarce, professional-grade content, becomes unlimited; what came as a bonus, the guarantee that an apparatus stands behind, becomes the decisive asset. The question every recipient will ask is no longer "is this work well done?", which they never knew how to judge, but "who stands behind it, and what do they risk?". In other words: value migrates from content to provenance. And provenance is exactly what the professions, alone, know how to produce.
When content costs nothing, provenance becomes the only thing that stays expensive.
The trade of trust
This reversal sketches the professions' future more surely than any speculation about their tools. Their apparent historical product, expert content, commoditizes. Their real historical product, the trust that spares verification, has never been worth more. The professions are not threatened by AI in what is essential to them; they are threatened with mistaking what, in them, is essential. Those that define themselves as producers of content will fight the unlimited, and lose. Those that define themselves as producers of trust will discover they hold the monopoly on the one thing the machine does not make: the institutional apparatus that allows believing without verifying.
This supposes work the professions have not yet begun: extending their trust machine to assisted work. Defining what the title guarantees when part of the work comes from a system; stating what conditions of supervision, governance and control authorize engaging the professional signal on a mixed production; sanctioning the engagement of the signal outside those conditions. It is an institutional undertaking, of the same order as the founding of the professional bodies in their time: once again, it is about building the apparatus that will allow society not to verify. The professions did it once, facing the asymmetry of knowledge. They must do it again, facing the asymmetry of the plausible. The technology of trust needs an update, and no one but them can write it.
The professions do not sell content. They sell the right not to verify. That right has never been worth more.
