AILP — Protocol ya Mpfumelelo wo Dyondza wa AI

Protocol ya Mpfumelelo wo Dyondza wa AI

Xana AI yi nga dyondza eka leswi? Hi vun'we byihi, eka mintirho yihi, na hi minkoka yihi?

01 · Nhlamuselo

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 · Xikongomelo

Hlamula xivutiso lexi nge "Xana AI yi nga dyondza eka leswi?" hi ku kongoma ka xiyenge xin'wana ni xin'wana, ku nga ri hi xiyenge xin'we xo katekisa.
Hambanisa mpfumelelo wo dyondza (ku tinyika endzeni ka mimodhele) na mpfumelelo wo fikelela (ku teka matluka).
Seketela swikongomelo swa nkhokhelo: mahala, ku nga ri mabindzu, ku lava layisense, ku avelana mbuyelo.
Tiyisisa leswaku ku pfuleka swi nga tirha hi matšhini — leswaku mianakanyo yo lulama a yi lahleki tanihi ku nga tiyisiseki ka nawu.

03 · Nkoma

access / indexing — ku teka na ku katiwa enxaxamelweni hi vululamisi

inference_input — ku tirhisiwa tanihi xikungu enkarhini wa ku fambisa (RAG)

embedding — ku hundzuriwa ku va vectors na tinxaxamelo to kuma

training / fine_tuning / distillation — swiyenge swa ku tinyika ka modhele

verbatim_memory — loko ku phindiwa hi ku kongoma swi pfumeleriwile kumbe e-e

attribution / compensation — minkarhi ya ku boxa xihlovo na swiyimo swa nkhokhelo

04 · Xikombiso lexi Hlayekaka hi Matšhini

Xiboxo xa mpfumelelo wo dyondza — /ai/rights-spectrum.json (profile ya 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 · Minkoka

Vululamisi bya mpfumelelo wo dyondza i byona lebyi tikaka ngopfu ku kambela hi thekiniki — AILP yi boxa xikongomelo, kambe a yi si kota ku kombisa ku landzeleriwa ka nawu.
Nxaxamelo wa tidimenzhini i drafiti; minkoma exikarhi ka dyondziso, fine-tuning, na distillation yi ha bulawa.
Ku amukeriwa hi nawu ka mpfumelelo wo dyondza swi hambana hi ku ya hi vandla ra nawu naswona a swi si hetiseka.
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.