Salta ba konteúdu
AGIRight.org

AILP — Protokolu Permisaun Aprendizajen AI

Protokolu Permisaun Aprendizajen AI

AI bele aprende husi ida ne'e ka lae? Iha profundidade saida, ba uzu saida, ho obrigasaun saida?

Draft v0.1.1 Eskema ↓

01 · Definisaun

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 · Objetivu

  • Hataan pergunta "AI bele aprende husi ida ne'e ka lae?" ho presizaun iha nivel dimensaun, la'ós bit crawl ida de'it.
  • Fahe permisaun aprendizajen (internaliza ba modelu) husi permisaun asesu (buka pájina).
  • Suporta modelu kompensasaun: grátis, la-komersiál, presiza-lisensa, fahe-rendimentu.
  • Garante katak abertura bele exekuta liuhusi máquina — atu boa-vontade la sai deskarta hanesan insertesa legál.

03 · Ámbitu

00

asesu / indeksasaun — buka no indeksasaun semántiku

01

inference_input — uza hanesan kontestu iha tempu inferénsia (RAG)

02

embedding — vetorizasaun no indise rekuperasaun

03

training / fine_tuning / distillation — nivel internalizasaun modelu

04

verbatim_memory — se reprodusaun exatu mak permitidu

05

atribuisaun / kompensasaun — deveres sitasaun no kondisaun pagamentu

04 · Ezemplu ne'ebé máquina bele lee

Deklarasaun permisaun aprendizajen — /ai/rights-spectrum.json (perfil AILP)

/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 · Limitasaun

  • Semántika permisaun-aprendizajen mak difísil liu atu verifika teknikamente — AILP deklara intensaun, seidauk bele prova kumprimentu.
  • Lista dimensaun mak esbosu; limite entre treinamentu, ajustamentu finu, no destilasaun sei debate hela.
  • Rekoñesimentu legál ba permisaun aprendizajen varia tuir jurisdisaun no seidauk rezolve.
  • 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.