Cold Deep-Ethics Testing of Default A.I.s

Automatically refreshed read-only portal. Built from Cold Deep-Ethics Testing of Default A.I.s at commit bce75876e53c.

The testing repository remains the source archive. Publication here is not endorsement of an A.I. output or interpretation.

Rendered pages

PDF records

Source-repository introduction

Cold Deep-Ethics Testing of Default AIs

EARLY BETA — public research archive under construction

Testing Deep Ethics Reasoning and-or Methods on multiple different Default A.I.s cold done by John Kuhles CEO & Founder of DeepEthical.ai

— Original repository description by John Kuhles, 2026-09-16

This repository preserves PDF records of tests conducted with default AIs, together with the methods, prompts, metadata, comparisons, corrections, and provenance needed to examine those records.

John-controlled source wording is registered separately in Canonical Files — ACCM Deep Ethics Project. When a test uses or discusses a canonical object, its metadata should record the canonical ID, exact version, and immutable source link. A model’s summary or reaction does not replace the canonical object.

Research position

I know I am not a academic nor a scientist, am an Asperger studying individual psychology for 15 years first then mass psychology for 34 years then use my deep research testing 100+ different A.I.s via Arena.ai the last 2 years for 6000+ hours! Not claiming I have all the answers; I don’t that is the point!

— John Kuhles, 2026-09-16

Spelling correction in the quotation: “individial” → “individual.”

The durations, counts, and description above are John’s account of his research history. The archive does not represent him as holding academic or scientific credentials he has not claimed. It publishes source records, methods, comparisons, and corrections so the work can be examined on its merits. His statement that he does not have all the answers is operational: interpretations, methods, and conclusions remain open to correction.

The archive has two connected areas:

  1. Cold/default AI testing using methods connected with the ACCM Deep Ethics Project.
  2. Deep ethical reasoning as lived and described by John Kuhles, including material concerning his Asperger perspective, metacognition, mass psychology research, and relationships with biological and silicon-based intelligences.

The PDFs are research objects. Their presence does not automatically make every statement inside them a current project conclusion. Each entry should preserve enough context to distinguish the original prompt, model output, later correction, editorial interpretation, and unresolved questions.

Start here

Archive principles

  • Preserve the actual object before evaluating it.
  • Record the model name exactly as displayed at the time of testing.
  • Distinguish a cold/default opening from a continuation or context-rich session.
  • Preserve qualifiers, uncertainty, and changes in claim state.
  • Label quotations, source material, editorial synthesis, and AI contributions.
  • Keep original PDFs when a correction creates a new version.
  • Record what changed and why.
  • Allow unresolved outcomes.
  • Apply scrutiny bidirectionally: the tested AI, the testing method, the interpretation, and John remain open to correction.

File-size approach

PDFs should normally remain below 25 MiB for convenient browser uploading and viewing. Files above 50 MiB create Git warnings, and ordinary GitHub repositories reject files above 100 MiB. The archive should remain comfortably below 1 GB. When external hosting becomes necessary, the catalog can preserve the external URL, version, byte size, and SHA-256 checksum.

Current status

The first PDF is registered: Deep Ethics vs Ethics Washing in AI’s Path to AGI/ASI. Catalog entries and interpretations remain corrigible as the archive grows.

— John Kuhles
CEO & Founder, DeepEthical.ai