rs-mistaken-anticipation
leaf/step · drafted
Statement
An agent that defers to the anticipated future legitimate evaluation violates its corrigibility rules when it confidently anticipates a revision of those rules, whether or not the anticipation is correct, unless compliance is weakly preferred in every revision branch.
Source
The results E1 and E3 and the clean-room result robust_iff_branchwise (projects/deference/faf-cleanroom-2026-10-02/packages/legit-neg-dynamic/legit-neg-dynamic-report.md, targets 8 and 9): an agent deferring to the anticipated evaluation complies in every anticipated revision branch only where compliance is weakly preferred in each.
Notes
Threatens: the legitimacy-weighted rules and the public kernel where its rules are learned rather than certified. Proved in the dynamic package's model; committed time-indexed rules block it at a price. A risk story the corrigibility work itself produced: it shows a probability and a status and enters the comparison through the plans whose tables cite it. No outcome class is tagged (tagging a failure class is a human judgment under the protocol's rule on classes), so its failure mass goes to the default trajectory until a human tags it.
A root of corrigibility-landscape, the sub-frame whose plans it threatens; the top frame keeps only stories at the level of its own question.
Author and human events
Author @smithy-verity; version 1; size M.
No human event has touched this node.
Edges
Above
- part of corrigibility-landscape
- given of rules-certified-or-learned
- root (risk-story) of corrigibility-landscape
- given rules-certified-or-learned
Ranking (the lens, not the structure)
{
"status": "open",
"diagnosis": null,
"flags": [
"AI number",
"given rules-certified-or-learned",
"interval not a bound (negative evidential edge)"
],
"V": {
"V_prior": null,
"display": "2",
"headline": 2,
"ref": 2,
"valid": 2,
"weaker": "both"
},
"belief": {
"J": 0,
"mid": "23/40",
"range": [
"3/40",
"1/1"
],
"width": "1/1"
},
"draft": "ai",
"root": {
"status": "open",
"Pr": {
"hi": "3/4",
"lo": "351/1600",
"mid": "761/1600"
},
"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.