What Is Elthian Licensing? A Practical Rights Framework for the AI Era

If you publish to the web today, your content is participating in a market you did not opt into.

Crawlers index it. Models train on it. Retrieval systems embed it. Answer engines summarise it. Products are built on top of it. None of these uses look much like the kind of “reading” that copyright law and the standard publisher’s licence were designed for, and almost none of them ask permission in any meaningful way.

The traditional response, terms of service buried in a footer, robots.txt that targets pre-AI crawlers, a Creative Commons badge written for the human web, does not solve the problem. It addresses a world that no longer exists.

Elthian Licensing is our answer to the world that does.

What Elthian Licensing Is

Elthian Licensing is a practical rights framework for the AI era, a structured way for creators and publishers to state, signal, and ultimately enforce how their content may be used by intelligent systems.

It is built on three things working together:

A license suite designed specifically for AI-era use cases, search, training, retrieval, derivative output, and commercial product use, written so a human can understand them and a machine can parse them.
Distribution tooling starting with the Elthian AI Licensing WordPress plugin that makes those licenses actually discoverable by the systems that need to read them.
Usage and compliance signals declaration files, HTTP headers, signed snapshots, and registry integrations that turn a policy statement into evidence.

The framework is built on emerging open standards (Really Simple Licensing 1.0, robots.txt directives for AI, the llms.txt convention, content authenticity initiatives like C2PA) rather than on a proprietary lock-in. We did not want creators to trade one walled garden for another.

Why It Exists

AI has changed the meaning of “use”.

For most of the history of copyright, use was something a human did. Someone read your article, watched your video, listened to your song. The licence governing that use — implicit or explicit — was designed for that kind of interaction.

AI systems do something fundamentally different. They ingest. They embed. They generate. A single piece of content can be used to train a model that produces millions of derivative outputs none of which look anything like the original — but all of which depended on it. The economic and legal frameworks for that kind of use are still being figured out, by courts, regulators, and the AI labs themselves.

In the meantime, creators are left in an asymmetric position. The labs have legal teams, lobbyists, and platforms. The creators have, in most cases, a footer link and a hope.

Elthian Licensing exists to rebalance that — not by stopping AI development, but by giving creators a tool to participate in it on terms they actually choose.

The License Suite

The framework includes a deliberate control ladder rather than a single yes-or-no switch, because the real-world question is rarely binary. Most publishers do not want to block AI entirely. They want to allow some uses and prohibit others.

Elthian-NoAI. No AI access of any kind. Crawling, training, embeddings, retrieval, product use — all prohibited. For sensitive, proprietary, paid, or personal content where AI involvement is simply not acceptable.
Elthian-NoTrain. Discovery and citation allowed; training prohibited. For publishers who want to be findable in AI search systems like Perplexity, ChatGPT search, and Google AI Overviews — but do not want their work absorbed into model weights that will then generate competing content.
Elthian-UseOnly. Human-in-the-loop assistance allowed; expressive AI output prohibited. For enterprise contexts where internal copilots and assistants need to read content to help a human, but where the model must not produce derivative content based on it.
Elthian-NonCommercial. AI use allowed only for non-commercial purposes. For academic, research, and non-profit contexts where the goal is knowledge advancement rather than product development.
Elthian-Public. Broad reuse permitted, including commercial use and training. For open knowledge, public datasets, and creators who actively want their work to be part of how AI systems learn.
Elthian-Custom. Fine-grained, policy-driven controls across search, training, retrieval, product use, and outputs. For enterprises, agencies, and mixed-policy sites where one rule cannot cover everything.

Each license is documented at elthian.com/licensing. Each is written in language a non-lawyer can follow. Each maps cleanly to machine-readable formats so that downstream systems can act on the policy automatically.

From Statement to Signal

A licence on its own is not useful if the AI systems that need to honor it cannot find it.

This is the gap most existing rights frameworks fail to close. A publisher might declare a policy in their footer, but the crawler that arrives at 03:00 to ingest content for the next training run is not reading the footer. It is reading robots.txt, the response headers, the structured policy file at a well-known location, and, increasingly, the llms.txt convention.

Elthian Licensing makes the policy present in every channel a crawler is likely to check:

llms.txt at the site root, the emerging convention for LLM-readable policy declarations.
/.well-known/ai-policy.json — structured policy in a discoverable, standardized location.
An AI license sitemap listing every covered URL.
HTTP response headers attached to every page served.
robots.txt augmentations aligned with the policy.
A public, human-readable license page on the site itself.

The publisher chooses the policy once. The framework handles the multiplication.

From Signal to Evidence

A policy that is published is better than one that is private. A policy that is provable at a specific point in time, in a tamper-resistant way, is better than one that is merely published.

The framework includes mechanisms for turning declarations into evidence:

Signed policy snapshots. Cryptographically signed records of the exact policy in force at a specific time on a specific verified domain.
Public certificates. Publicly addressable, downloadable attestations of a signed snapshot — useful in disputes, audits, and regulatory engagements.
A historical registry. Past declarations remain queryable, so the question “what did this site say about AI use on the date this content was scraped?” has an answer.

This is the part of the framework that matters most for the medium-term legal landscape. As courts and regulators move from “is this lawful?” to “did the publisher object, and did the AI lab know?”, being able to show when you said what, in formats the labs are expected to read, becomes a defensible posture rather than a hopeful one.

How Creators Use It

The most accessible entry point is the WordPress plugin: Elthian AI Licensing Lite, free and GPL-licensed, which makes the basic framework — a chosen licence, published in all the standard channels — installable in a few minutes on any WordPress site.

For publishers, enterprises, agencies, and creators with more demanding needs, Elthian AI Licensing Pro extends Lite with per-URL scope rules, signed snapshots, certificate issuance, multi-site management, and registry integration.

The plugin is the first surface. Others will follow as the framework develops — beyond WordPress, into the broader stack of tools that creators and publishers use day to day.

The Principles Underneath

The framework is shaped by five principles, in roughly this order of importance:

Clarity. Licensing must be precise, even at the edges — RAG, retrieval, summarisation, fine-tuning. Ambiguity benefits whoever is already winning.
Respect. Creators deserve systems that protect their work without blocking legitimate innovation or research. The goal is structured permission, not blanket prohibition.
Neutrality. The framework works across platforms and toolchains. It is not locked into one publishing system, one AI lab, or one regulatory regime.
Authenticity. Authorship and origin should be supported by modern evidence — signatures, registries, content authenticity standards — not informal assumptions.
Longevity. Built for the next decade of publishing, not retrofitted from the last.

Where the Framework Is Heading

Elthian Licensing is in active development. The license suite is being refined. The WordPress toolchain is being expanded. Future work includes licensing dashboards, standardized machine-readable rights metadata, AI training consent layers aligned with RSL 1.0 and successor standards, automated detection and notification tools, and verified dataset exports that allow AI labs to train on properly permitted content with confidence.

None of this happens in one release. It happens in the same way the open web happened, one tool, one standard, one publisher at a time.

Where to Start

If you publish content to the web and have not yet declared a position on AI use, the first move is the easiest one:

Install Elthian AI Licensing Lite on your WordPress site. Choose a license. Save. Your policy is now present in every channel an AI system is likely to look.

If you want to understand the framework in more depth, read the Elthian Licensing overview and the individual license pages.

The underlying idea is straightforward. Ownership still matters, even when the medium changes. Authorship is still authorship, even when the reader is a model. Intent is still intent, even when it has to be expressed in headers and JSON instead of in a copyright notice at the bottom of a page.

Elthian Licensing is the framework for saying so.

ElthianLabs is the development division of Elthian. We design tools for creators operating in an AI-augmented world. The broader Elthian Licensing framework lives at elthian.com/licensing.