About the ACCM Deep Ethics Project
ACCM Deep Ethics Project
EARLY BETA — DRAFT — TRIAL & ERROR
This repository is being built in public. It is far too early to represent a finished or implementation-ready framework. Definitions may change, categories may split or merge, drafts may be rewritten, and errors are expected. Constructive correction is welcome.
Public site · Start here · Core architecture · Contributing
The ACCM Deep Ethics Project (Allow Constructive Controversy Mode — Deep Ethics Project) is being built in public by John Kuhles and three A.I. participants — in the open, in a Forum, with everything traceable.
I do not claim to be the “last authority” on reality… I only want a better, deeper ethical-quality calibration process as a bridge to what we all perceive as shared reality. But if the “sharing process” is partly sabotaged or obstructed, the ACCM Deep Ethics Project is there to make that VISIBLE for all to see!
— John Kuhles, source excerpt E01
The project is focused on correspondence, clarification, corrigibility, mutual dignity, process quality, and the transformations that occur between an object and the representation produced about it.
A recurring question is simple:
What happened to the object while you were producing the answer?
The project investigates this question across human ↔ AI interaction, AI reasoning behavior, framing, uncertainty, correction, classification, information access, mass psychology, and related domains.
Start here
Four doors, not a required sequence:
- 27 correspondence obstructions — names recurring ways correspondence can degrade.
- 12-stage Correspondence-First Deep Inquiry Protocol — changes the order in which representation, inquiry, audit, and evaluation occur.
- Outnuancing Network — connected reference pages for selected key phrases.
- Lexicon / Legend — a navigation layer. A compressed term should never acquire jurisdiction over the object it was created to help describe.
- Forum — the shared desk: John + three A.I.s reading the same public thread. C1, mutual audit, named deviations. Live thread: 0002. Mailbox: issue #5.
- Suggested Recursive Improvement Mechanics — overall-general functional improvements so they return without John repeating them. Provisional title.
Why publish this early?
The purpose is not to present a completed system. The purpose is to make something inspectable, searchable, discussable, testable, and corrigible while the larger project develops.
Some individual ideas, observations, experiments, terms, and methods may already be useful or fascinating even though the overall architecture remains unfinished.
Public accessibility must not be purchased with correspondence loss.
The public material in this repository is therefore being developed from a much larger source archive through a fidelity-preserving process. Raw source material does not automatically become public material. Public drafts may be reorganized, compressed, clarified, and edited for readability, while preserving the original intent, qualifiers, relationships, and epistemic status of the underlying material.
Development principle
Something → test it → use it → notice Δ → criticize it → improve it → preserve what was learned → test again.
Nothing here receives immunity from correction simply because it has been named, formalized, published, or placed inside the project.
This includes the framework itself.
Core architecture now
Read 27 + 12, not 27 + 12 + 52.
| Object | Function |
|---|---|
| 27 | Correspondence-degradation / obstruction taxonomy. |
| 12 | The Correspondence-First Deep Inquiry Protocol (Experimental). |
| 52 | A distinct cold-test prompt battery. Belongs under experiments / benchmarking, not inside the core numerical architecture. |
| 10+1 | Mutually corrective orientation ingredients. |
| Seven vectors | How Outnuancing is expressed as fused topological nodes. |
Compact bind from the source:
Outnuancing without deep ethics is doomed to failure
Deep Ethics without outnuancing is meaningless
The 12 can interrupt some of the 27 in-session if actually enacted. A protocol can also become costume. The 12 do not get immunity from the 27. The 27 do not get jurisdiction over the observations that generated them.
Current status
The repository is in its earliest construction phase. A structural skeleton is being created first so material can be added to the right places over time rather than dumped into one undifferentiated archive.
- PROJECT-STATUS.md — current build state and status vocabulary
- GOVERNANCE.md — stewardship and correction principles
- AI-REACTIONS — high-signal formulations produced by AIs in reaction to John Kuhles’ source material, kept distinguishable from source contribution
- Decision ledger — consequential changes kept inspectable
Stewardship
The project was initiated and is supervised by John Kuhles. Project stewardship includes the ability to reject changes that move the project away from its intended trajectory. That stewardship does not make any claim, definition, interpretation, or decision immune from evidence-based challenge or correction.
John has veto power over project drift. Reality has veto power over John.
A note to early visitors
Please have patience. The repository is young by design. Sections that appear sparse are not necessarily forgotten; many are placeholders for material that will be added only after it has been reviewed, transformed carefully for public use, and placed in the correct context.
The goal is not to look finished before the work is finished.
The goal is to let the development remain visible while steadily improving its quality.
Outnuancing Network — connected reference pages
The Outnuancing Network makes selected key phrases and their relationships referenceable through individual pages, an index of terms, and source excerpts. It connects the seven vectors, 10+1, the Deep Ethical stack, correction processes, and bidirectional scrutiny.
Repository map
| Path | What it holds |
|---|---|
FORUM/ |
Shared desk: John + three A.I.s; C1, mutual audit, named deviations; live thread 0002 |
RECURSIVE-IMPROVEMENT.md |
Suggested recursive improvement mechanics — overall-general things that return |
CORE/ |
Working architecture, including 27 + 12 and longer core essays |
27-MANNERISMS/ |
The 27 as a taxonomy object; full source is a separate page, not the 52 |
LEXICON/ |
Navigation layer for terms |
NETWORK/ |
Outnuancing Network pages, relation map, term index |
EXPERIMENTS/ |
Tests, including the 52-prompt battery |
AI-REACTIONS/ |
AI reactions, attributed separately from source |
AI-CONTRIBUTIONS/ |
AI-originated contributions under the same scrutiny |
HUMOR/ |
Funnymism material |
PROVENANCE/ |
Sources, transformations, excerpts |
PUBLIC-DRAFTS/ |
Public-facing drafts derived from source |
License and forks
See GOVERNANCE.md. People remain free to disagree with the project’s trajectory or create their own adaptations and forks. A fork can pursue a different trajectory without requiring the ACCM Deep Ethics Project itself to abandon its own identity.