AI content licensing for publishers in Asia

1. What AI content licensing is, and why now

AI content licensing defines how participating AI companies, search platforms, agents and enterprise systems may access and use your content, under agreed permissions and commercial terms. Revenue depends on buyer demand, the rights offered and the applicable agreement.

The open web was built for people to search, click and visit. The agentic web is being built for AI search, crawlers, agents and enterprise systems that read content on a reader's behalf, often without sending a visit back. When fewer readers click through, the old model of earning from advertising on page views weakens.

Licensing gives publishers a commercial path into that new environment, ideally before the terms are set by platforms alone.

2. What you can license

Two questions define a licence: what content is included, and which uses are allowed.

Content

Paywalled content can be included or excluded. You can also set scope at the level of domains, subdomains, sections or content types, so a licence never has to cover everything you publish.

Uses

• Training: content used to build or improve AI models.
• Grounding and retrieval: content fetched at the moment an AI system answers a question, so the answer reflects current, reliable information.
• AI search and agents: content summarized, quoted or acted on inside AI search results, assistants and agent workflows.

Each use carries a different value and risk, so many publishers allow some uses while requiring a separate licence, or refusing, for others.

3. Three licensing routes compared

Once your scope and rights are clear, there are three main ways to take content to market.

These routes are not mutually exclusive. A publisher might enroll its archive in a collective, list a data product on a marketplace, and pursue a direct deal for real-time news. Before combining routes, check how each agreement's permissions, restrictions and obligations overlap for the same content. Exclusivity is one of these, but not the only one.

LINC is a local operating partner in Asia for one collective route, the Internet Licensing Collective (ILC).

4. How pricing is changing: from pay-per-crawl to pay-per-use

Pricing for AI access to content is shifting. The first models charged each time a crawler fetched a page; newer models pay when content is actually used in an AI answer. As of October 2026, what is still missing is a transparent, shared way to value that use.

Pay-per-crawl, introduced by Cloudflare in July 2025, lets a site set one price per request and charge AI crawlers each time they fetch a page. It is simple and the publisher sets the price, but it pays for access rather than value: Cloudflare has noted that much legitimate crawler traffic re-fetches pages that have not changed.

Pay-per-use pays when content is actually used to produce an AI answer or result. Three versions are emerging:

• Collective licensing (ILC): AI companies pay when members' content is used to generate output, under rates and royalty rules set collectively for all members.
• Cloudflare Pay Per Use: announced in July 2026 and opened as a beta in late September 2026. The AI buyer proposes the paid use and the rate, the publisher chooses whether to accept, and usage is reported by the buyer. Cloudflare says publishers can choose between this and pay-per-crawl.
• Google AI contribution pilot: a Search Console pilot that pays some publishers when their content significantly contributes to answers in Google's AI features. Google determines contribution and payment, and participants have said they cannot see how it is calculated.

Use the same four questions for any pay-per-use program you are offered, and check whether participation limits your other routes. Who sets the rate is not the same as transparency: also ask how usage is measured and whether it can be verified. LINC helps publishers compare these models against their rights, goals and existing agreements.

Sources: Cloudflare on pay-per-crawl and Cloudflare (September 2026) on Pay Per Use; TechCrunch (July 2026); Digiday and Search Engine Roundtable on Google's pilot; the RSL standard.

5. How licensing terms are expressed

Contracts set the commercial terms, but AI systems also need a machine-readable way to see them. The RSL (Really Simple Licensing) standard lets you publish licensing terms in an XML file and associate them with your content through robots.txt, HTTP headers, web pages, RSS feeds or media files, so terms can differ by path, page, feed or asset; the RSL specification covers the details. Keep in mind that robots.txt communicates instructions to crawlers but does not itself enforce access restrictions (see the robots.txt standard, RFC 9309).

6. What to prepare

You don't need everything below before your first conversation, but the clearer these points are, the faster you can evaluate the routes.

1. A rights inventorySeparate what you own outright from what you license in: wire and agency copy, freelance and contributor work, syndicated articles, stock and agency photography, and user-generated content. You can usually only license what you hold the rights to.
2. Existing AI and syndication agreementsList every current AI licence or content deal, and note any exclusivity or field-of-use restrictions. New scope has to be defined around them.
3. Content scopeDomains, subdomains and sections; archive depth; paywalled areas; formats such as text, images, video, audio and data; and languages.
4. ExclusionsContent you never want licensed, or only licensed for certain uses.
5. Access points and metadataFeeds, APIs, sitemaps and archives, plus consistent metadata such as dates, authors, sections and language. Clean metadata makes content easier to license and report on.
6. A picture of current AI accessWhich AI crawlers already visit your sites, how often, and what your robots.txt currently allows.
7. A decision-maker and domain controlSomeone who can approve scope on the publisher's behalf, and control of the domains you want to enroll, so licensing declarations can be published.

This guide is general information, not legal advice. Copyright rules for AI use differ between markets and are still changing, so take local legal advice on your specific rights and agreements.

7. Licensing across Asia

Asia is not one content market. Publishers here face conditions that make licensing harder to do alone:

• Many languages: AI companies need high-quality content in each market's language, and a catalog's value often depends on how well it covers a language or region.
• Fragmented markets: buyers want scale, but each market has its own publishers, rights practices and commercial norms.
• Frameworks built elsewhere: many licensing standards and collectives were developed in Western markets and need local onboarding, technical implementation and representation to work for Asian publishers.
• Changing rules: how copyright applies to AI use is still being decided in many markets, which makes clear rights records and flexible, non-exclusive arrangements especially useful.

8. Where LINC fits

LINC is an AI content licensing network for publishers in Asia, operated by IntentBridges in Singapore with LINC K.K. in Japan. It handles the publisher-side work in three stages:

LINC helps publishers manage the work across these routes, maintaining continuity as content, permissions and commercial relationships develop.

• Non-exclusive: joining LINC does not itself require exclusivity. Each licensing opportunity is assessed against your existing agreements.
• Your approval first: registration does not authorize content licensing. Content scope, permissions and the applicable approval process are agreed separately.
• Free to register: registering interest is free and creates no commitment. Any service fees and commercial terms are agreed before the relevant work begins.