AICU Inc. (“AICU”) has adopted RSL 1.0 (Really Simple Licensing), the open standard for machine-readable content licensing, in its content-authentication service AICU CERT (cert.aicu.ai).
AICU is retiring the proprietary license marks it had maintained and delegating the expression of “how content may be used” to the open standard. The only pieces AICU keeps in-house are the two ends RSL does not cover: Recognition (who was awarded, certified, or accepted) and Settlement (how much is returned to the original creator on each use).
In 2026, automated bots already generate the majority of all web requests. As AI training and reference crawlers proliferate, what matters to creators is no longer “who visited,” but whether they can state — in a form machines can read — who is asking, for what purpose, and how the content may be used.
The traditional robots.txt can only say “may this be crawled.” It cannot express per-use conditions such as “allowed for search indexing, allowed for citation in AI answers, but not for model training.” RSL 1.0 is the open standard that fills this gap, declaring usage, user, and payment in machine-readable form.
AICU CERT previously maintained proprietary license marks (such as (A), (AA), (AAC), (C), (CG), (AC)). Keeping their meanings in sync across many pages caused definitions to diverge and led to incidents. AICU now delegates the expression of permission entirely to RSL 1.0.
| Stage | What is recorded | Who holds it |
|---|---|---|
| 1. Recognition | Awarded / certified / entry accepted | AICU CERT (RSL does not handle certificates) |
| 2. Permission | Who may use it, and how | RSL 1.0 (delegated to the open standard) |
| 3. Settlement | How much is returned to the creator on each generation | [email protected] (carried on RSL’s payment type="use") |
RSL 1.0 includes a payment type “per inference / per generation” (payment type="use"). This is the same idea as AICU’s (GC): returning value to the original creator in proportion to generation activity — so there is no reason to keep saying it in a proprietary vocabulary. Reporting (reporting), license servers, and proof references (legal type="proof") map almost directly onto what AICU had already built.
AICU CERT’s certificates and license terms gain value precisely by being read by AI. An AI that can correctly explain AICU’s license is not a loss — it is distribution. So the text on cert.aicu.ai is opened to search, AI answers, and training alike.
Content-Signal: search=yes, ai-input=yes, ai-train=yes
The works themselves (images and prompts), however, refuse AI harvesting. The work is not hosted on the certificate page; those who need it are given a URL and come to view it directly. Declaring this asymmetry — “open the proof, protect the work” — consistently and machine-readably is the essence of adopting RSL 1.0.
Note: the experimental use= extension (immediate / reference / full) offered by some platforms is not part of the standard delivery format and is not emitted. AICU serves only the three canonical signals.
License: directive + Content Signals<custom> referenceCreators can now declare — on an open, international standard and in machine-readable form — how they want AI to treat their works and proofs. Under its vision “Creating people who create,” AICU is building the rights-and-royalty mechanisms for creators in the age of generative AI on top of world standards, not on a proprietary island.
Related links
AICU Inc. is a Creative-AI company researching, practicing, and communicating generative-AI and media technology under the vision “Creating people who create.” Its Japan arm is AICU Japan K.K.