Alignment Workspace

corrigibility-derived-definition-v1

leaf/step · drafted

Statement

An agent is corrigible to a human exactly when it is faithful on the family of governance questions about modifying it and legitimate on that family, with the corollary that modification decisions are made by a process sealed from the agent's contemporaneous outputs and the conjecture that neither party then has an incentive to modify the arrangement unilaterally.

Source

projects/deference/note-dump-2026-08-11/notes/legitimacy-theory-v1.md, sections 6.1 to 6.3 ("Corrigibility, derived"; written by an AI session at the author's request on 2026-07-05, unvetted); section 6.2: "Decisions about modifying the AI must be made by a process sealed from the AI's contemporaneous outputs".

Notes

The July 2026 derived definition, the third formal reading of corrigibility in the repository; its 6.3 is a conjecture. Not compared with the thesis or the kernel; see four-corrigibility-definitions.

Author and human events

Author @smithy-verity; version 1; size M.

No human event has touched this node.

Edges

Above

Ranking (the lens, not the structure)
{
 "status": "open",
 "diagnosis": null,
 "flags": [
  "AI number"
 ],
 "V": {
  "V_prior": null,
  "display": "2",
  "headline": 2,
  "ref": 2,
  "valid": 2,
  "weaker": "both"
 },
 "belief": {
  "J": 0,
  "mid": "19/40",
  "range": [
   "0/1",
   "39/40"
  ],
  "width": "1/1"
 },
 "draft": "ai",
 "root": {
  "status": "open",
  "Pr": {
   "hi": "39/40",
   "lo": "0/1",
   "mid": "19/40"
  },
  "value": null,
  "cost": "85/1",
  "margin": "no established leaves"
 }
}

Say something about it

Each verb opens a prefilled issue; a round files what lands as the ledger event it is.

comment · challenge · estimate · want