Spring na inhoud
AGIRight.org

AILP — Protokol vir KI-leertoestemming

Protokol vir KI-leertoestemming

Mag KI hieruit leer? Tot watter diepte, vir watter gebruike, met watter verpligtinge?

Draft v0.1.1 Skema ↓

01 · Definisie

AILP operationalizes the AIRS spectrum for the specific question of learning. It distinguishes acts that binary crawler rules collapse together: being read is not being learned from; being retrieved is not being trained on; fine-tuning is not distillation; verbatim memorization is different from statistical internalization. AILP lets a publisher declare, in a single machine-readable file, per-dimension permissions — access, indexing, inference input, embedding, training, fine-tuning, distillation, memory, output, attribution and compensation — each as allowed, denied, or license-required. An AILP grant answers "may AI learn from this content?" only — it does not answer "may this Agent invoke a tool, write to a database, or publish?" (AARS), nor "who authorized this Agent?" (AADP).

02 · Doel

  • Beantwoord die vraag „mag KI hieruit leer?” met presisie op dimensievlak, in plaas van 'n enkele crawl-bit.
  • Skei leertoestemming (internalisering in modelle) van toegangstoestemming (die ophaal van bladsye).
  • Ondersteun vergoedingsmodelle: gratis, nie-kommersieel, lisensie-vereis, inkomstedeling.
  • Verseker dat openheid masjien-uitvoerbaar is — sodat goeie trou nie as regsonsekerheid weggegooi word nie.

03 · Omvang

00

access / indexing — ophaal en semantiese indeksering

01

inference_input — gebruik as konteks tydens inferensie (RAG)

02

embedding — vektorisering en herwinningsindekse

03

training / fine_tuning / distillation — vlakke van model-internalisering

04

verbatim_memory — of presiese reproduksie toegelaat word

05

attribution / compensation — aanhalingspligte en betalingsvoorwaardes

04 · Masjienleesbare voorbeeld

'n Verklaring van leertoestemming — /ai/rights-spectrum.json (AILP-profiel)

/ai/rights-spectrum.json
{
  "version": "0.1",
  "protocol": "AILP",
  "publisher": "example.org",
  "default": {
    "access": "allowed",
    "indexing": "allowed",
    "inference_input": "allowed",
    "embedding": "allowed",
    "training": "license_required",
    "fine_tuning": "license_required",
    "distillation": "denied",
    "verbatim_memory": "denied",
    "attribution": "required",
    "compensation": "contact"
  },
  "contact": "licensing@example.org"
}

05 · Beperkings

  • Die semantiek van leertoestemming is tegnies die moeilikste om te verifieer — AILP verklaar voorneme, maar kan nog nie nakoming bewys nie.
  • Die lys van dimensies is 'n konsep; grense tussen afrigting, fyninstelling, en distillasie word steeds bespreek.
  • Regserkenning van leertoestemmings verskil van jurisdiksie tot jurisdiksie en is onopgelos.
  • v0.1.1: AILP(Resource,ContentUse)=Allow does not imply AARS(Actor,Action)=Allow, and does not imply AADP(Principal,Actor,Authority)=Valid — a system exposing both content and tools must evaluate them separately.