How a contocracy works

Contocracy is a form of rule proposed in 2026. One chosen person holds the community’s highest office for good. Lower powers may be delegated, but the top office is not transferable and passes to no heir. On the contocrat’s death it is held in abeyance and remains hers. Between death and any return, decisions that can wait do wait, and operators she chose in life judge which matters are urgent. One declared AI stand-in acts on those urgent matters by default, or for a living contocrat who cannot be reached, if authorised. Everyone knows the stand-in, not the contocrat, is deciding, and its rulings are labelled as its own. Her religion prevails over the mission. Cryopreservation is expected but not required. No one has been revived, and the form promises nothing. A proven return, which takes proof of first-person continuity (a likeness does not qualify), would bring a full briefing and the power to reverse the stand-in’s rulings.

Two rules turn her record into infrastructure. The stand-in may act only on what her lifelong record supports, and a rebuilt or upgraded model is allowed only if it matches the contocrat’s record at least as well as the model before it. The record is therefore the evidence behind every interim ruling and the yardstick for every new model. That puts four familiar duties under new strain.

What the record demands

Retention comes first. Most retention schedules end in disposal, and this one never does. It works like a legal hold, which pauses deletion for litigation, but with no release date. It covers drafts, email, meeting notes, marginalia and voice memos as well as final decisions. Appraisal shifts from choosing what to destroy to choosing what to describe first. Every gap has a price, because each silence in the record narrows what the stand-in can support. Her religious commitments, which outrank the mission, should be documented in her own words.

Provenance, the trail showing where each item came from and who made it, matters as much. Okafor must know which words are truly hers. A letter she signed but a deputy drafted would teach a model someone else’s judgment, so authorship has to be recorded item by item. The stand-in’s rulings belong in the archive too, but in a separate series under its own label, so they never pass as hers and never shape what the next model learns.

Integrity is where authority is most exposed. If the record is the yardstick, anyone who can quietly alter it can steer the stand-in or rig an upgrade test. The defence is ordinary preservation practice applied without exceptions: fixity checks that reveal any change to a file, several copies in separate hands (the principle behind LOCKSS), and a log of every access. Her own record closes at her death, so a sealed, checksummed copy can serve as the fixed reference for every later test.

Access needs layers. The stand-in reads everything, and operators see enough to judge urgency. Members need enough to check any ruling, so each ruling should cite the records it relies on. People who wrote to her keep their privacy, so public copies will need redaction while the originals stay whole. Because a proven return brings a full briefing, the stand-in’s labelled rulings and their citations also become the briefing file.

The upgrade rule then works like a migration test. Preservation teams judge a file-format migration by whether it keeps the significant properties of the original. Contocracy asks the same of judgment. Before any change, the current stand-in and the candidate face the same cases drawn from the record, such as documented decisions with her reasoning attached. Each is scored on how closely it matches what she actually decided, and the candidate takes over only if it matches at least as well.

The strongest objections

The first objection is privacy. Under European rules, personal data must be stored for the shortest time possible, with exceptions such as archiving in the public interest, and her files are full of other people. The tension is real. A community adopting contocracy would need a lawful basis for each kind of third-party data, consent where it can be had, and firm redaction rules. The second is completeness. No record captures a whole person, and much of what anyone knows never reaches a file. That is why the stand-in acts only where the record supports it, only on urgent matters and only under its own label. It never claims to be her.

The third objection is capture, since whoever controls the archive could steer the stand-in. Separate custody, published fixity reports and cited rulings are the check, and the operators she chose in life decide which matters reach the stand-in at all. The fourth is cost. Preservation with no end date is expensive, yet it is the field’s daily work, minus the end date.

What records professionals can do now

If communities adopt contocracy, records professionals will build what it runs on, long before any death. A contocrat’s record needs a retention plan with no end date, item-level authorship, copies in separate custody, layered access with redaction and a published method for the upgrade test. I invite records managers, archivists, librarians and digital preservation specialists to treat contocracy as a design brief and to draft the standards a contocrat’s record would need, before anyone has to rely on them.

Disclosure: Claude Opus 5.5 (Anthropic) was used to draft the text and the figures from my concept and to check the sources it cites; I reviewed and approved the final text. Competing interests: I proposed the term contocracy. Funding: none.