The jury of peers

There is a strange, almost wasteful extravagance at the heart of the jury. To decide whether one person is guilty of a crime, a functioning legal system will stop everything, summon a dozen strangers who know nothing about the case and are chosen precisely for that ignorance, seat them in a box, walk them slowly through the evidence, and then send them into a room to argue until they agree. These are not experts. They are not, as a rule, trained in law or in the weighing of proof. They are borrowed for the occasion from ordinary life and returned to it when the verdict is read. On any measure of speed or expertise, this is a preposterous way to reach a conclusion. And yet the deepest tradition in the common law insisted that for the gravest judgments — those that could take a person's liberty or life — this was the only legitimate way to reach one.

The instinct the jury answers is not epistemic but political. It is not, in the first instance, a claim that twelve laypeople will get the facts more right than a single learned judge. It is a claim about who may be trusted to pass judgment at all. And it is very old. The question the jury settles is one every society that concentrates power has to answer sooner or later, and it is worth stating plainly before turning to how the jury answered it: when the stakes are highest, by whom, and by what process, may a person legitimately be judged?

Judged by peers, not power

The English jury did not arrive fully formed; it accreted over centuries, out of older customs of sworn neighbours and inquest, and only slowly became the fact-finding body of laypeople we now recognise. But its charter moment — the phrase that fixed the principle in the political imagination — comes from Magna Carta in 1215. Its most famous clause promises that no free man shall be seized, imprisoned, dispossessed, outlawed, or exiled "except by the lawful judgment of his peers or by the law of the land." The barons who forced that concession from King John were not drafting a theory of evidence. They were trying to put a wall between a person and the arbitrary judgment of the crown. The point of judicium parium — judgment by peers — was that grave verdicts should not issue from the king or his officers alone, but from a body of the accused's own equals.

Over the following centuries that guarantee matured into the criminal jury: a group of ordinary citizens, drawn from the community, charged with deciding the facts and, in a criminal case, the question of guilt. What is striking is that this was a deliberate choice against the more efficient alternatives. A single official could have decided faster. A trained magistrate could have decided, in the narrow sense, more competently. The crown, in most of the systems the common law competed with, was perfectly happy to decide for itself. The jury was the road not taken toward centralised judgment — an insistence that the power to condemn be lodged not in a person or an office but in a plurality of peers who had no stake in the outcome and would go home afterward to live among the consequences of what they had decided.

The jury was never an argument that ordinary people know more. It was an argument that some judgments are too grave to be made by power alone, however capable.

Legitimacy from structure

Why locate the highest-stakes judgments in a plurality of independent, deliberating peers rather than in the most competent available decider? Because the jury encodes a claim that is easy to miss and hard to overstate: for the gravest judgments, legitimacy is served not by expertise or efficiency but by structure. The authority of a jury's verdict does not come from any juror being wise. It comes from the arrangement — several independent laypeople, none able to decide alone, required to deliberate together, to reach their verdict collectively, and to do so publicly enough that the judgment can be seen and held to account. That structure is doing the legitimating work. It disperses the power to condemn so that no single will can exercise it; it forces the reasons into the open of a shared room; it grounds the verdict in the community that must live with it.

This is a specific and unusual answer to the problem of grave judgment. It says that how a judgment is reached, and by whom, can matter as much as whether the judgment is correct — that a verdict handed down by a single unaccountable official, even a right one, carries a different and thinner authority than the same verdict reached by a jury of the affected person's peers. The jury is, in this sense, a machine for manufacturing legitimacy out of structure: collective rather than singular, independent rather than directed, of the community rather than of the state.

None of this requires pretending the jury is a good instrument for finding truth. It is slow, expensive, and fallible. Juries have convicted the innocent and freed the guilty, and carried the prejudices of their communities into the box. The institution's flaws are real and not marginal, and it would be dishonest to invoke the jury as a model of accuracy. The point is not that the jury reliably gets the facts right. The point is what the institution was reaching for — a settled answer to the question of who may legitimately judge, and by what process, when the stakes are at their highest. It located that authority in structure precisely because it could not be located, safely, in any single person's competence or good faith.

By whom may a person be judged

Now consider the shape of a consequential automated judgment — the kind that decides a benefit, a loan, a flag, a sentence-adjacent score. On nearly every axis, it is the jury's exact inverse. It is a single system rather than a plurality. It is opaque rather than deliberating in the open. There is no collective, no argument in a shared room, no body of peers, no community that must live among the consequences of what it decided. There is, most often, nothing that confers legitimacy on the judgment beyond the bare fact that the system was authorised to make it. The decision is fast, it is confident, and it is structurally solitary. It answers the question by whom may a person be judged with a shrug: by whatever was deployed.

The jury's quiet argument is that this is not enough — that for the gravest judgments, being authorised and even being correct do not exhaust what legitimacy requires. Legitimacy, on the jury's account, is a property of the structure through which a judgment is reached, not only of its authority or its accuracy. Some decisions are too consequential to rest on a single unaccountable decider, however capable that decider is. And machine judgment at scale reopens, in a form the barons at Runnymede could not have pictured, the precise question the jury was built to answer: when a system's verdict can take something grave from a person, by whom, and by what process, may that person legitimately be judged?

I want to be careful about the shape of this claim, because it is an argument by analogy and analogies can be pushed past their strength. I am not saying every automated decision needs a literal jury, or that software should be adjudicated by a dozen citizens in a box. That would misread what the jury was for. The jury is not a template to copy; it is a demonstration that a civilisation, having watched grave judgment concentrate in single hands, once decided that the legitimacy of such judgment must come from its structure — collective, independent, accountable, of the affected person's community — and built an institution to make that true. What the public tools of accountability can borrow is not the jury box but the principle behind it: that a Decision Receipt, standing, contestability, the second opinion, and clear provenance are not courtesies bolted onto an authorised decision but the beginnings of a structure through which a machine judgment might earn a legitimacy that mere authorisation can never confer. The jury asked who may judge, and how. We are being asked it again, at a volume the old answer never anticipated, and we should not pretend the question is new.

— Dispatches · Summit Cognitive

---