Ucwaningo

Amalungelo, imithwalo, izibopho, kanye nokulawulwa kwe-AI, i-AGI, kanye nama-ejenti — kusuka emathuluzini kuya kubasebenzisani kuya kubantu abangaba khona esikhathini esizayo. Lokhu kufaka umbuzo wokuvikelwa kokuziphatha okuncane: yimiphi imithetho okufanele ilawule ukuxhumana phakathi kwabantu ne-AI ngaphambi kokuthi imibuzo yobuntu ixazululwe.

What AI systems may do with content: read, summarize, transform, retrieve, train, commercialize, redistribute. The AICR ruleset and the AICL-C content-licensing layer make these rights declarable and transactable.

Ukuthi ngabe i-AI ingafunda, ngobunzulu obungakanani, ngezinjongo ezini, futhi ngaphansi kwemithwalo enjani. Ukufundwa akufani nokufundelwa kukho; "akuvinjelwe" akufani "nokufundeleka". I-AIRS ne-AILP ziguqula imvume yokufunda ibe isipektramu esinyukayo, esifundeka yimishini.

Indlela ama-ejenti afinyelela ngayo amawebhusayithi, ama-API, izinkokhelo zedatha, izisekelo zolwazi, kanye nokuqukethwe okukhokhelwayo: ubunikazi, imvume, izicelo, inkokhelo, igunya, amarekhodi okusetshenziswa, imikhawulo yezokuphepha, kanye nesivikelo sokuhlaselwa kwe-prompt-injection.

What an Agent may DO once it holds tool or API capability, as distinct from what it may read or learn: action taxonomy, effect vectors, reversibility, blast radius, and multi-action composition risk. AARS gives content permission and action permission separate, independently-evaluable axes.

Who an Agent acts for, where its authority came from, whether it can hand that authority to another Agent, and how much of its internal state a verifier may demand. AADP separates principal from actor, bounds delegation to strictly narrow (never silently widen), and treats inspection as a bounded right, not an unlimited one.

Imithetho yokulawulwa engatholwa futhi yenziwe yimishini: i-llms.txt, imanifesti ye-/ai/, amafayela enqubomgomo e-/.well-known/, ama-JSON Schema, amathokheni elayisensi asayiniwe, kanye namafomethi amarekhodi okucwaningwa. Le sayithi uqobo iyisibonelo esisebenzayo.

Umnotho-zombangazwe we-AI nokuqukethwe: i-pay-per-crawl, amadivhidendi edatha, izikhwama ze-AI ezibusayo, amaqoqo enkokhelo yabadali, imakethe zedatha ezingamazinga, kanye nendlela inani elikhishwe olwazini lomphakathi elingabuyela ngayo kulabo abalikhiqizile.