AI Creator Licensing Rights: Your Work, Your Terms, Your Price
The PM said creators control the price. But what does a meaningful permission system actually look like? Here’s the architecture your rights need beneath them.
Ashe Davis
Founder & CEO, ScriptShield
Ashe Davis is the founder of ScriptShield. She is an optioned Screenwriter, Author, and SaaS Developer, who wanted to help creators prove authorship and protect their intellectual property in the digital age.
Three words cut through everything the Prime Minister said on Tuesday: control the price.
Not “receive compensation.” Not “participate in a licensing framework.” Control the price. That means you — the person who wrote the book, composed the song, designed the artwork — decide what your work is worth when an AI company comes asking. ScriptShield was built around this idea before the government caught up to it.
But here is the question nobody in Canberra answered: what does that control actually look like as a system? If a creator can set the price, where is it recorded? Who verifies the terms? What happens when they are breached?
The policy is historic. The infrastructure beneath it does not exist yet. This article is about what that infrastructure must contain — and what you can start building today.
The Short Answer: AI creator licensing rights require more than an opt-out button. A meaningful permission system records which work, which use, which company, which model, what price, what duration and what conditions apply. ScriptShield’s AI Training Rights Passport is designed to hold exactly these terms. Protect your work today.
What did the PM mean by “control the price”?
The speech was specific. Anthony Albanese said no company should use Australian books, music, art or news to build or train AI without the artist’s control. He then added the clause that matters most: “That includes the artist’s control of the price and value of their work.”
Annabelle Herd, CEO of ARIA, put it in commercial terms: “Control of price, value and terms of use are what underpin a commercial licensing market. The artist decides what their work is worth and who may use it. That is how licensing works everywhere else in the world and it is how it should work here.”
This is not an abstract idea. It is how music licensing already works. APRA AMCOS manages performing rights. Screenrights manages retransmission. Copyright Agency manages reprographic rights. Each one operates a system where creators set terms, licensees pay and records verify the arrangement.
AI training needs the same thing — but at a different scale, with different granularity and without an established collective framework to manage it. That is the gap.
Why can’t creators just say “yes” or “no” and be done?
Because “yes” and “no” do not describe a licensing relationship.
Imagine a published novelist. A major AI company approaches her about including her backlist in a training dataset. Here are the questions she needs answered before she can make an informed decision.
Which books? All twelve, or just the three that are out of print? Which company? Is the licence transferable to subsidiaries or partners? Which model? A general-purpose language model, a specialised literary analysis tool, or both? Which purpose? Pre-training, fine-tuning, retrieval-augmented generation, synthetic output? For how long? Twelve months, three years, perpetual? In which territory? Australia only, or worldwide? With attribution? Her name attached to outputs derived from her work? At what price? Per title, per model, per query, per year? Can the permission be revoked? What happens to a model already trained on her work?
A checkbox handles none of this. A properly structured permission record handles all of it.
What should a creator permission system actually contain?
This is the architecture ScriptShield is building inside the Rights Graph.
The protected work
Identified by its SHA-256 cryptographic hash, RFC 3161 trusted timestamp and position in its timestamped version chain. Not “your content” — the specific file, edition and version.
The rights holder
Verified through the account identity and, where available, through linked publication records, ISBNs, registration numbers or publisher agreements. Chain of title documentation establishes who has the authority to grant permission.
The permission profile
Your AI Training Rights Passport records separate conditions for each type of use: pre-training, fine-tuning, retrieval-augmented generation, embeddings, search indexing, model evaluation, synthetic-data generation, commercial output and non-commercial research.
The commercial terms
Price, payment method, payment status, invoicing and settlement. If the PM says you control the price, the system must record what you set and whether it was paid.
The boundaries
Duration, territory, attribution requirements, sublicensing restrictions and revocation conditions. Every permission must have edges. Open-ended grants create disputes.
The AI-use declaration
Your three-tier declaration records whether the work itself involved AI — because an AI company licensing human-authored work has different obligations than one licensing AI-assisted work.
The receipt
When permission is granted and terms are accepted, the system should issue a signed, timestamped, tamper-evident receipt connecting the protected work, the rights holder, the permitted use, the price and the receiving company. That receipt becomes evidence for both parties.
What standards exist to support machine-readable licensing?
The technical landscape is maturing quickly. Several standards are moving towards the kind of granular, machine-readable licensing that the PM’s vision requires.
Really Simple Licensing (RSL), launched in September 2025, goes further than any other standard. It supports specific AI use cases — training, inference, fine-tuning — and specific payment types including subscription, per-crawl and per-inference. Tim O’Reilly called it “the missing licensing layer for the AI-first Internet.” It is early-stage but pointed in the right direction.
C2PA content credentials can now embed training and data-mining assertions directly into files. The CAWG Training and Data Mining Assertion allows creators to signal whether an asset may be used for data mining, AI inference, model training or generative training — each set to allowed, not allowed or constrained.
W3C ODRL provides a formal vocabulary for expressing permissions, prohibitions and obligations around digital content. It is designed for exactly this kind of structured rights expression.
ScriptShield’s approach is to interoperate with these standards rather than replace them. RSL can express the terms. C2PA can carry the signal. ODRL can structure the policy. ScriptShield connects them to the work’s provenance record, the creator’s identity, the version history and the Rights Graph — creating a persistent, verifiable, creator-controlled record that follows the work through its lifecycle.
What can creators do right now, before the licensing framework arrives?
Do not wait. The creators who document their position now will be the ones who negotiate effectively when the framework arrives.
Set your terms
Use your AI Training Rights Passport to record your licensing position for each type of AI use. Even without a formal licensing mechanism, a documented, timestamped position is evidence of your intent.
Protect your catalogue
Every work you plan to licence should have a provenance baseline — hash, timestamp, version chain, authorship declaration. Start with a free hash receipt and build from there.
Declare your authorship
Your AI & Human Authorship declaration establishes whether AI was involved in creating the work. That matters for licensing because human-authored and AI-assisted works carry different rights implications.
Keep records of every approach
If an AI company, platform or publisher contacts you about licensing, record the correspondence, the proposed terms and your response. Attach it to the relevant work in your Rights Graph. Future disputes often turn on what was offered and what was agreed.
Engage with the process
The CAIRG is still exploring what Australia’s licensing model will look like. The Senate inquiry into AI and data centres is still accepting submissions. Creator voices shape the infrastructure.
The Bottom Line
The Prime Minister gave creators the most powerful words any government has spoken about AI and creative rights: your work, your terms, your price. Those words are correct. They are also, right now, words without machinery.
ScriptShield is building the machinery. The Rights Graph connects provenance to permission to price to payment to ongoing compliance. The AI Training Rights Passport records your terms before anyone asks. The evidence record follows your work wherever it travels.
Set Your Terms Today
Start with a free hash receipt. Record your AI Training Rights Passport position. Because when the framework arrives, documented creators negotiate — and undocumented creators accept someone else’s terms.
Get Your Free Hash ReceiptProtect Your Creative Work
Generate SHA-256 authorship certificates and track who sees your scripts, manuscripts, and creative works.
Get StartedScriptShield provides evidence documentation tools for creators. It is not a law firm and does not provide legal advice.